
import pandas as pd

from prophet import Prophet
from prophet.plot import plot_plotly
from prophet.plot import add_changepoints_to_plot

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import plotly.offline as py
py.init_notebook_mode()
%matplotlib inline
import seaborn as sns
from pylab import rcParams

# To supress warnings
import warnings
warnings.filterwarnings("ignore")


#We are providing inputs to tell pandas that we are trying to work with time series.
df = pd.read_csv(r'C:\APZIVA-2023\Project-5\valueinvestor-project5.csv',  parse_dates = True)


df.columns,

(Index(['Date', 'Price', 'Open', 'High', 'Low', 'Vol.', 'Change %'], dtype='object'),)


df.shape

(312, 7)


#Check data types
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 312 entries, 0 to 311
Data columns (total 7 columns):
 #   Column    Non-Null Count  Dtype 
---  ------    --------------  ----- 
 0   Date      312 non-null    object
 1   Price     312 non-null    object
 2   Open      312 non-null    object
 3   High      312 non-null    object
 4   Low       312 non-null    object
 5   Vol.      311 non-null    object
 6   Change %  311 non-null    object
dtypes: object(7)
memory usage: 17.2+ KB


df.head()


# Assuming the DataFrame has been imported and named 'df'

# Delete the last row as it is providing statistics
df = df[:-1]

# Reverse the order so that 2020 appears first
df = df[::-1]

# Reset the index
df = df.reset_index(drop=True)

# Drop the 'Vol.' and 'Change %' columns
df = df.drop(columns=['Vol.', 'Change %'])

# Checking for null values and dropping them if present
if df.isnull().values.any():
    df = df.dropna()

# Display the cleaned DataFrame
print(df)

             Date   Price    Open    High     Low
0    Jan 03, 2020     255  255.99  258.19   253.7
1    Jan 06, 2020   253.9  254.75  254.84   251.4
2    Jan 08, 2020  259.15  253.57  259.15  253.03
3    Jan 09, 2020  257.99   259.4  261.76  257.01
4    Jan 10, 2020  258.19  257.86  259.25  256.88
..            ...     ...     ...     ...     ...
306  Mar 25, 2021  286.66   286.5  287.03  283.85
307  Mar 26, 2021  291.22  288.63  292.75  288.32
308  Mar 29, 2021   293.3  289.72  294.09  289.26
309  Mar 30, 2021  293.25  294.86  295.72   291.5
310  Mar 31, 2021  291.02     294  294.42  290.26

[311 rows x 5 columns]


df.index.dtype

dtype('int64')


# Check for missing values
print(df.isnull().sum())

Date     0
Price    0
Open     0
High     0
Low      0
dtype: int64


df.columns

Index(['Date', 'Price', 'Open', 'High', 'Low'], dtype='object')


# We have considered Price as a dependent variable for this session. Readers are advised to look at  'Vol.','open' price as dependent variable for univariate Time Series Prediction
df_date = df[['Date','Price']]


import pandas as pd
from pandas.tseries.offsets import BDay  # Import BDay from pandas.tseries.offsets

# Generate a date range for business days from January 1, 2020, to March 31, 2021
date1 = pd.date_range(start='1/3/2020', end='3/31/2021', freq=BDay())


# Display the generated date range
print(date1)

DatetimeIndex(['2020-01-03', '2020-01-06', '2020-01-07', '2020-01-08',
               '2020-01-09', '2020-01-10', '2020-01-13', '2020-01-14',
               '2020-01-15', '2020-01-16',
               ...
               '2021-03-18', '2021-03-19', '2021-03-22', '2021-03-23',
               '2021-03-24', '2021-03-25', '2021-03-26', '2021-03-29',
               '2021-03-30', '2021-03-31'],
              dtype='datetime64[ns]', length=324, freq='B')


from pandas.tseries.holiday import USFederalHolidayCalendar
from pandas.tseries.offsets import CustomBusinessDay
us_bd = CustomBusinessDay(calendar=USFederalHolidayCalendar())
us_bd

<CustomBusinessDay>


# Generate a date range for business days from January 1, 2020, to March 31, 2021
date = pd.date_range(start='1/3/2020', end='3/31/2021', freq=us_bd)


date

DatetimeIndex(['2020-01-03', '2020-01-06', '2020-01-07', '2020-01-08',
               '2020-01-09', '2020-01-10', '2020-01-13', '2020-01-14',
               '2020-01-15', '2020-01-16',
               ...
               '2021-03-18', '2021-03-19', '2021-03-22', '2021-03-23',
               '2021-03-24', '2021-03-25', '2021-03-26', '2021-03-29',
               '2021-03-30', '2021-03-31'],
              dtype='datetime64[ns]', length=312, freq='C')


df_date['Time_Stamp'] = pd.DataFrame(date,columns=['Date'])
df_date.head()
df_date.tail()


import matplotlib.pyplot as plt

# Set the figure size
plt.rcParams['figure.figsize'] = 15, 8


import pandas as pd

# Assuming df_date['Time_Stamp'] is already in datetime format
df_date['Price'] = pd.to_numeric(df_date['Price'], errors='coerce')

# Using .loc to set the 'Price' column
df_date.loc[:, 'Price'] = pd.to_numeric(df_date['Price'], errors='coerce')


# Plotting
plt.plot(df_date['Time_Stamp'], df_date['Price'])
plt.grid()
plt.show()


df_date.describe()


## Alternate way of reading the data as time series.
## Squeese = True will return the index col as series

df_final = pd.read_csv(r'C:\APZIVA-2023\Project-5\valueinvestor-project5.csv', parse_dates=True, index_col=0)
df_final = df_final.squeeze()


df_final.head()


df_final.info()

<class 'pandas.core.frame.DataFrame'>
Index: 312 entries, Mar 31, 2021 to Highest: 296.07
Data columns (total 6 columns):
 #   Column    Non-Null Count  Dtype 
---  ------    --------------  ----- 
 0   Price     312 non-null    object
 1   Open      312 non-null    object
 2   High      312 non-null    object
 3   Low       312 non-null    object
 4   Vol.      311 non-null    object
 5   Change %  311 non-null    object
dtypes: object(6)
memory usage: 17.1+ KB


# Convert numeric columns to float
numeric_columns = ['Price', 'Open', 'High', 'Low', 'Vol.', 'Change %']

for col in numeric_columns:
    df_final[col] = pd.to_numeric(df_final[col], errors='coerce')

# Check the data types again
print(df_final.dtypes)

Price       float64
Open        float64
High        float64
Low         float64
Vol.        float64
Change %    float64
dtype: object


rcParams['figure.figsize'] = 15,8


df_final.plot();
plt.grid()


df_final['Price'].plot();
plt.grid()


# Convert the index to DateTimeIndex
df_final.index = pd.to_datetime(df_final.index, errors='coerce')

# Check if the conversion was successful
print(type(df_final.index))

# Create the boxplot
sns.boxplot(x=df_final.index.year, y=df_final['Open'])
plt.grid()
plt.show()

<class 'pandas.core.indexes.datetimes.DatetimeIndex'>


sns.boxplot(x = df_final.index.month_name(),y = df_final['Open'])
plt.grid();


df_date.head()
df_date.tail()


# Decompose the 'Volume' variable Time Series.

df_date['Time_Stamp'] = pd.to_datetime(df_date['Time_Stamp'])
df_model = df_date.set_index('Time_Stamp')
df_model.drop(['Date'], axis=1, inplace=True)
df_model.tail()


from    statsmodels.tsa.seasonal import   seasonal_decompose


from statsmodels.tsa.seasonal import seasonal_decompose
import matplotlib.pyplot as plt

# Drop missing values in 'Price' column
df_model.dropna(subset=['Price'], inplace=True)

# Convert 'Price' column to numeric values
df_model['Price'] = pd.to_numeric(df_model['Price'], errors='coerce')

# Perform seasonal decomposition
decomposition = seasonal_decompose(df_model['Price'], model='additive', period=5)

# Plot the decomposition
fig, (ax1, ax2, ax3, ax4) = plt.subplots(4, 1, figsize=(10, 8), sharex=True)
decomposition.trend.plot(ax=ax1, title='Trend')
decomposition.seasonal.plot(ax=ax2, title='Seasonal')
decomposition.resid.plot(ax=ax3, title='Residual')
df_model['Price'].plot(ax=ax4, title='Original')

plt.tight_layout()
plt.show()


trend = decomposition.trend.fillna(value=0)
seasonality = decomposition.seasonal.fillna(value=0)
residual = decomposition.resid.fillna(value=0)

print('Trend','\n',trend.head(12),'\n')
print('Seasonality','\n',seasonality.head(12),'\n')
print('Residual','\n',residual.head(12),'\n')

Trend 
 Time_Stamp
2020-01-03      0.000
2020-01-06      0.000
2020-01-07    256.846
2020-01-08    258.326
2020-01-09    259.356
2020-01-10    258.788
2020-01-13    258.692
2020-01-14    259.554
2020-01-15    260.330
2020-01-16    262.132
2020-01-17    264.178
2020-01-21    265.422
Name: trend, dtype: float64 

Seasonality 
 Time_Stamp
2020-01-03   -0.014090
2020-01-06   -0.076582
2020-01-07   -0.091924
2020-01-08   -0.167150
2020-01-09    0.349746
2020-01-10   -0.014090
2020-01-13   -0.076582
2020-01-14   -0.091924
2020-01-15   -0.167150
2020-01-16    0.349746
2020-01-17   -0.014090
2020-01-21   -0.076582
Name: seasonal, dtype: float64 

Residual 
 Time_Stamp
2020-01-03    0.000000
2020-01-06    0.000000
2020-01-07    2.395924
2020-01-08   -0.168850
2020-01-09   -1.515746
2020-01-10    3.626090
2020-01-13    0.434582
2020-01-14   -3.152076
2020-01-15   -2.652850
2020-01-16    0.018254
2020-01-17    2.116090
2020-01-21    2.714582
Name: resid, dtype: float64


from statsmodels.tsa.seasonal import seasonal_decompose
import matplotlib.pyplot as plt

# Assuming 'df_model' is your DataFrame with a DateTimeIndex
decomposition = seasonal_decompose(df_model['Price'], model='multiplicative', period=5)

# Plot the decomposition
fig, (ax1, ax2, ax3, ax4) = plt.subplots(4, 1, figsize=(10, 8), sharex=True)
ax1.plot(decomposition.trend, label='Trend')
ax1.set_title('Trend')
ax2.plot(decomposition.seasonal, label='Seasonal')
ax2.set_title('Seasonal')
ax3.plot(decomposition.resid, label='Residual')
ax3.set_title('Residual')
ax4.plot(df_model['Price'], label='Original')
ax4.set_title('Original')

plt.tight_layout()
plt.show()


trend = decomposition.trend.fillna(value=0)
seasonality = decomposition.seasonal.fillna(value=0)
residual = decomposition.resid.fillna(value=0)

print('Trend','\n',trend.head(12),'\n')
print('Seasonality','\n',seasonality.head(12),'\n')
print('Residual','\n',residual.head(12),'\n')

Trend 
 Time_Stamp
2020-01-03      0.000
2020-01-06      0.000
2020-01-07    256.846
2020-01-08    258.326
2020-01-09    259.356
2020-01-10    258.788
2020-01-13    258.692
2020-01-14    259.554
2020-01-15    260.330
2020-01-16    262.132
2020-01-17    264.178
2020-01-21    265.422
Name: trend, dtype: float64 

Seasonality 
 Time_Stamp
2020-01-03    0.999551
2020-01-06    0.999722
2020-01-07    0.999952
2020-01-08    0.999430
2020-01-09    1.001344
2020-01-10    0.999551
2020-01-13    0.999722
2020-01-14    0.999952
2020-01-15    0.999430
2020-01-16    1.001344
2020-01-17    0.999551
2020-01-21    0.999722
Name: seasonal, dtype: float64 

Residual 
 Time_Stamp
2020-01-03    0.000000
2020-01-06    0.000000
2020-01-07    1.009019
2020-01-08    0.999268
2020-01-09    0.994168
2020-01-10    1.014412
2020-01-13    1.001663
2020-01-14    0.987549
2020-01-15    0.989731
2020-01-16    1.000060
2020-01-17    1.008409
2020-01-21    1.010220
Name: resid, dtype: float64


### We will consider the original dataset and subseqeuntly we would require exogeneous variables

df['Time_Stamp'] = pd.DataFrame(date,columns=['Date'])
df.head()
df.tail()


df['Time_Stamp'] = pd.to_datetime(df['Time_Stamp'])
df_final_model = df.set_index('Time_Stamp')
# df_final_model.drop(['Date'], axis=1, inplace=True)
df_final_model


# Split the series into train and test
df['Time_Stamp'] = pd.to_datetime(df['Time_Stamp'])
df_final_model['Month'] = df_final_model.index.month
df_final_model['Year'] = df_final_model.index.year
df_final_model.head()


df_final_model.tail()


df.tail()


train = df_final_model[pd.to_datetime(df_final_model['Date']) < pd.to_datetime('2021-01-31')]
train.shape

(269, 7)


test = df_final_model[pd.to_datetime(df_final_model['Date']) >= pd.to_datetime('2021-01-31')]
test.shape

(42, 7)


train_final = train[['Price']]
test_final=test[['Price']]


print('Training Data')
display(train)
print('Test Data')
display(test)

Training Data

Test Data


print(train_final.shape)
print(test_final.shape)

(269, 1)
(42, 1)


train_final.tail()


test_final.tail()


print('First few rows of Training Data')
display(train.head())
print('Last few rows of Training Data')
display(train.tail())
print('First few rows of Test Data')
display(test.head())
print('Last few rows of Test Data')
display(test.tail())

First few rows of Training Data

Last few rows of Training Data

First few rows of Test Data

Last few rows of Test Data


print(train.shape)
print(test.shape)

(269, 7)
(42, 7)


## Test for stationarity of the series - Dicky Fuller test

from statsmodels.tsa.stattools import adfuller
def test_stationarity(timeseries):
    
    #Determing rolling statistics
    rolmean = timeseries.rolling(window=7).mean()
    rolstd = timeseries.rolling(window=7).std()

    #Plot rolling statistics:
    orig = plt.plot(timeseries, color='blue',label='Original')
    mean = plt.plot(rolmean, color='red', label='Rolling Mean')
    std = plt.plot(rolstd, color='black', label = 'Rolling Std')
    plt.legend(loc='best')
    plt.title('Rolling Mean & Standard Deviation')
    plt.show(block=False)
    
    #Perform Dickey-Fuller test:
    print ('Results of Dickey-Fuller Test:')
    dftest = adfuller(timeseries, autolag='AIC')
    dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','#Lags Used','Number of Observations Used'])
    for key,value in dftest[4].items():
        dfoutput['Critical Value (%s)'%key] = value
    print (dfoutput,'\n')


test_stationarity(train_final['Price'])

Results of Dickey-Fuller Test:
Test Statistic                  -1.678662
p-value                          0.442139
#Lags Used                       8.000000
Number of Observations Used    260.000000
Critical Value (1%)             -3.455754
Critical Value (5%)             -2.872721
Critical Value (10%)            -2.572728
dtype: float64


import itertools
p = q = range(0, 3)
d= range(0,1)
pdq = list(itertools.product(p, d, q))

model_pdq = [(x[0], x[1], x[2], 5) for x in list(itertools.product(p, d, q))]
print('Examples of parameter combinations for Model...')
print('Model: {}{}'.format(pdq[1], model_pdq[1]))
print('Model: {}{}'.format(pdq[1], model_pdq[2]))
print('Model: {}{}'.format(pdq[2], model_pdq[3]))
print('Model: {}{}'.format(pdq[2], model_pdq[4]))

Examples of parameter combinations for Model...
Model: (0, 0, 1)(0, 0, 1, 5)
Model: (0, 0, 1)(0, 0, 2, 5)
Model: (0, 0, 2)(1, 0, 0, 5)
Model: (0, 0, 2)(1, 0, 1, 5)


## The following loop helps us in getting a combination of different parameters of p and q in the range of 0 and 2
## We have kept the value of d as 1 as we need to take a difference of the series to make it stationary.

import itertools
p = q = range(0, 4)
d= range(1,2)
pdq = list(itertools.product(p, d, q))
print('Examples of the parameter combinations for the Model')
for i in range(0,len(pdq)):
    print('Model: {}'.format(pdq[i]))

Examples of the parameter combinations for the Model
Model: (0, 1, 0)
Model: (0, 1, 1)
Model: (0, 1, 2)
Model: (0, 1, 3)
Model: (1, 1, 0)
Model: (1, 1, 1)
Model: (1, 1, 2)
Model: (1, 1, 3)
Model: (2, 1, 0)
Model: (2, 1, 1)
Model: (2, 1, 2)
Model: (2, 1, 3)
Model: (3, 1, 0)
Model: (3, 1, 1)
Model: (3, 1, 2)
Model: (3, 1, 3)


# Creating an empty Dataframe with column names only
ARIMA_AIC = pd.DataFrame(columns=['param', 'AIC'])
ARIMA_AIC


import numpy as np
import pandas as pd
from statsmodels.tsa.arima.model import ARIMA

# Define the range for p, d, and q
p = range(0, 3)
d = range(0, 3)
q = range(0, 3)

# Create a list of tuples containing all possible combinations of p, d, and q
pdq = list(itertools.product(p, d, q))

# Initialize an empty list to store AIC values
ARIMA_AIC = []

# Replace this with your actual price data
price_data = [10, 20, 30, 40, 50, 60, 70, 80, 90, 100]

for param in pdq:
    ARIMA_model = ARIMA(price_data, order=param).fit()  # Replace this with your actual price data
    print('ARIMA{} - AIC:{}'.format(param, ARIMA_model.aic))
    ARIMA_AIC.append({'param': param, 'AIC': ARIMA_model.aic})

# Convert the list of AIC values into a DataFrame
ARIMA_AIC_df = pd.DataFrame(ARIMA_AIC)

ARIMA(0, 0, 0) - AIC:99.53260453230253
ARIMA(0, 0, 1) - AIC:92.94455876901922
ARIMA(0, 0, 2) - AIC:87.12385604775815
ARIMA(0, 1, 0) - AIC:68.98742527404838
ARIMA(0, 1, 1) - AIC:61.761697307249555

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

ARIMA(0, 1, 2) - AIC:55.66460899061261
ARIMA(0, 2, 0) - AIC:-160.74266169207124
ARIMA(0, 2, 1) - AIC:-158.74266169207124
ARIMA(0, 2, 2) - AIC:-156.74266169207124
ARIMA(1, 0, 0) - AIC:83.24107366580836
ARIMA(1, 0, 1) - AIC:76.22723219692597
ARIMA(1, 0, 2) - AIC:70.28715995871397
ARIMA(1, 1, 0) - AIC:-21.000326833341763

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

ARIMA(1, 1, 1) - AIC:-41.297878329222534
ARIMA(1, 1, 2) - AIC:-37.25240258194637
ARIMA(1, 2, 0) - AIC:-158.74266169207124

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

ARIMA(1, 2, 1) - AIC:-156.74266169207124
ARIMA(1, 2, 2) - AIC:-154.74266169207124
ARIMA(2, 0, 0) - AIC:-9.510482052122835

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

ARIMA(2, 0, 1) - AIC:-0.5022912376760083

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals


from statsmodels.tsa.arima.model import ARIMA
import pandas as pd
import itertools

# Load your dataset
train = pd.read_csv(r'C:\APZIVA-2023\Project-5\valueinvestor-project5.csv')

# Initialize an empty list to store AIC values
aic_values = []

# Define the p, d, and q values for ARIMA
p = range(3)
d = range(2)
q = range(3)
pdq = list(itertools.product(p, d, q))

# Price Data
train = pd.read_csv(r'C:\APZIVA-2023\Project-5\valueinvestor-project5.csv')

# Drop non-numeric values
train['Price'] = pd.to_numeric(train['Price'], errors='coerce')
train = train.dropna(subset=['Price'])

for param in pdq:
    ARIMA_model = ARIMA(train['Price'].values, order=param).fit()
    aic_values.append({'param': param, 'AIC': ARIMA_model.aic})

# Create a DataFrame from the list of dictionaries
ARIMA_AIC = pd.DataFrame(aic_values)

# Ensure the ARIMA_AIC DataFrame is sorted before using the head method
ARIMA_AIC.sort_values(by='AIC', ascending=True, inplace=True)
ARIMA_AIC.head()

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals


import pandas as pd
import numpy as np

# Load your dataset
train = pd.read_csv(r'C:\APZIVA-2023\Project-5\valueinvestor-project5.csv')

# Remove any rows where the 'Price' column contains non-numeric values
train = train[pd.to_numeric(train['Price'], errors='coerce').notna()]

# Convert the 'Price' column to float
train['Price'] = pd.to_numeric(train['Price'])

# Now, you should be able to run your ARIMA model

# Define the ARIMA model
auto_ARIMA = ARIMA(train['Price'], order=(2, 1, 2))

# Fit the ARIMA model to your data
results_auto_ARIMA = auto_ARIMA.fit()

# Print the summary of the model
print(results_auto_ARIMA.summary())

                               SARIMAX Results                                
==============================================================================
Dep. Variable:                  Price   No. Observations:                  311
Model:                 ARIMA(2, 1, 2)   Log Likelihood                -918.615
Date:                Wed, 21 Feb 2024   AIC                           1847.229
Time:                        10:48:05   BIC                           1865.912
Sample:                             0   HQIC                          1854.698
                                - 311                                         
Covariance Type:                  opg                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
ar.L1         -0.5590      0.019    -30.136      0.000      -0.595      -0.523
ar.L2         -0.9572      0.017    -56.128      0.000      -0.991      -0.924
ma.L1          0.6139      0.025     24.124      0.000       0.564       0.664
ma.L2          0.9981      0.078     12.873      0.000       0.846       1.150
sigma2        21.6907      2.046     10.602      0.000      17.681      25.701
===================================================================================
Ljung-Box (L1) (Q):                   0.10   Jarque-Bera (JB):               108.24
Prob(Q):                              0.75   Prob(JB):                         0.00
Heteroskedasticity (H):               1.68   Skew:                             0.30
Prob(H) (two-sided):                  0.01   Kurtosis:                         5.83
===================================================================================

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals


results_auto_ARIMA.plot_diagnostics();


predicted_auto_ARIMA = results_auto_ARIMA.forecast(steps=len(test))


from sklearn.metrics import mean_squared_error
from statsmodels.tsa.arima.model import ARIMA
import pandas as pd
import numpy as np

# Assuming 'test' contains the actual prices and 'predicted_auto_ARIMA' contains the predicted prices

test['Price'] = test['Price'].astype(float)

# Calculate RMSE
rmse = np.sqrt(mean_squared_error(test['Price'], predicted_auto_ARIMA))

# Calculate MAPE
def mean_absolute_percentage_error(y_true, y_pred):
    y_true, y_pred = np.array(y_true), np.array(y_pred)
    return np.mean(np.abs((y_true - y_pred) / y_true)) * 100

mape = mean_absolute_percentage_error(test['Price'], predicted_auto_ARIMA)

print('RMSE:', rmse)
print('MAPE:', mape)

RMSE: 24.676528039946877
MAPE: 8.14194236670457


resultsDf = pd.DataFrame({'RMSE': rmse,'MAPE':mape}
                           ,index=['ARIMA(2,1,2)'])

resultsDf


from statsmodels.graphics.tsaplots import plot_acf, plot_pacf
import matplotlib.pyplot as plt

# Plot autocorrelation
plot_acf(train['Price'].diff().dropna(), title='Training Data Autocorrelation')
plt.show()

# Plot partial autocorrelation
plot_pacf(train['Price'].diff().dropna(), title='Training Data Partial Autocorrelation')
plt.show()


import pandas as pd
import numpy as np

# Load your dataset
train = pd.read_csv(r'C:\APZIVA-2023\Project-5\valueinvestor-project5.csv')

# Remove any rows where the 'Price' column contains non-numeric values
train = train[pd.to_numeric(train['Price'], errors='coerce').notna()]

# Convert the 'Price' column to float
train['Price'] = pd.to_numeric(train['Price'])

# Now, you should be able to run your ARIMA model

# Define the ARIMA model
manual_ARIMA = ARIMA(train['Price'], order=(3, 1, 3))

# Fit the ARIMA model to your data
results_manual_ARIMA = manual_ARIMA.fit()

# Print the summary of the model
print(results_manual_ARIMA.summary())

                               SARIMAX Results                                
==============================================================================
Dep. Variable:                  Price   No. Observations:                  311
Model:                 ARIMA(3, 1, 3)   Log Likelihood                -921.684
Date:                Wed, 21 Feb 2024   AIC                           1857.368
Time:                        10:48:07   BIC                           1883.524
Sample:                             0   HQIC                          1867.824
                                - 311                                         
Covariance Type:                  opg                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
ar.L1          0.2053      0.203      1.011      0.312      -0.193       0.604
ar.L2         -0.2060      0.167     -1.233      0.217      -0.533       0.121
ar.L3          0.9285      0.212      4.376      0.000       0.513       1.344
ma.L1         -0.1744      0.219     -0.795      0.426      -0.604       0.255
ma.L2          0.2232      0.180      1.243      0.214      -0.129       0.575
ma.L3         -0.9633      0.241     -3.995      0.000      -1.436      -0.491
sigma2        22.0148      1.598     13.779      0.000      18.883      25.146
===================================================================================
Ljung-Box (L1) (Q):                   0.00   Jarque-Bera (JB):               125.03
Prob(Q):                              1.00   Prob(JB):                         0.00
Heteroskedasticity (H):               1.76   Skew:                             0.31
Prob(H) (two-sided):                  0.00   Kurtosis:                         6.05
===================================================================================

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals


results_manual_ARIMA.plot_diagnostics();


predicted_manual_ARIMA = results_manual_ARIMA.forecast(steps=len(test))


rmse = mean_squared_error(test['Price'],predicted_manual_ARIMA,squared=False)
mape = mean_absolute_percentage_error(test['Price'],predicted_manual_ARIMA)
print('RMSE:',rmse,'\nMAPE:',mape)

RMSE: 23.750883249627755 
MAPE: 7.844811065198802


temp_resultsDf = pd.DataFrame({'RMSE': rmse,'MAPE':mape}
                           ,index=['ARIMA(3,1,3)'])


resultsDf = pd.concat([resultsDf,temp_resultsDf])

resultsDf


from statsmodels.graphics.tsaplots import plot_acf

# Select 'Price' column and compute differenced series
price_diff = train['Price'].diff().dropna()

# Plot the autocorrelation of the differenced 'Price' series
plot_acf(price_diff, title='Training Data Autocorrelation')


import itertools
p = q = range(0, 4)
d= range(1,2)
D = range(0,1)
pdq = list(itertools.product(p, d, q))
PDQ = [(x[0], x[1], x[2], 6) for x in list(itertools.product(p, D, q))]
print('Examples of the parameter combinations for the Model are')
for i in range(1,len(pdq)):
    print('Model: {}{}'.format(pdq[i], PDQ[i]))

Examples of the parameter combinations for the Model are
Model: (0, 1, 1)(0, 0, 1, 6)
Model: (0, 1, 2)(0, 0, 2, 6)
Model: (0, 1, 3)(0, 0, 3, 6)
Model: (1, 1, 0)(1, 0, 0, 6)
Model: (1, 1, 1)(1, 0, 1, 6)
Model: (1, 1, 2)(1, 0, 2, 6)
Model: (1, 1, 3)(1, 0, 3, 6)
Model: (2, 1, 0)(2, 0, 0, 6)
Model: (2, 1, 1)(2, 0, 1, 6)
Model: (2, 1, 2)(2, 0, 2, 6)
Model: (2, 1, 3)(2, 0, 3, 6)
Model: (3, 1, 0)(3, 0, 0, 6)
Model: (3, 1, 1)(3, 0, 1, 6)
Model: (3, 1, 2)(3, 0, 2, 6)
Model: (3, 1, 3)(3, 0, 3, 6)


SARIMA_AIC = pd.DataFrame(columns=['param','seasonal', 'AIC'])
SARIMA_AIC


import pandas as pd
import statsmodels.api as sm

# Define ranges for SARIMA parameters
pdq = [(0,0,0), (0,0,1), (0,1,0), (0,1,1), (1,0,0), (1,0,1), (1,1,0), (1,1,1)]
PDQ = [(0,0,0,12), (0,0,1,12), (0,1,0,12), (0,1,1,12), (1,0,0,12), (1,0,1,12), (1,1,0,12), (1,1,1,12)]

# Initialize a list to store SARIMA AIC values
SARIMA_AIC = []

for param in pdq:
    for param_seasonal in PDQ:
        try:
            SARIMA_model = sm.tsa.statespace.SARIMAX(train['Price'],
                                            order=param,
                                            seasonal_order=param_seasonal,
                                            enforce_stationarity=False,
                                            enforce_invertibility=False)
            
            results_SARIMA = SARIMA_model.fit(maxiter=1000)
            print('SARIMA{}x{} - AIC:{}'.format(param, param_seasonal, results_SARIMA.aic))
            SARIMA_AIC.append({'param':param, 'param_seasonal': param_seasonal, 'AIC': results_SARIMA.aic})
        except:
            print('Error with SARIMA{}x{}'.format(param, param_seasonal))

# Convert the SARIMA_AIC list of dictionaries to a DataFrame for easy sorting
SARIMA_AIC_df = pd.DataFrame(SARIMA_AIC)

# Sort the DataFrame by AIC values in ascending order
SARIMA_AIC_df_sorted = SARIMA_AIC_df.sort_values(by='AIC')

# Retrieve the best SARIMA model parameters with the lowest AIC
best_SARIMA_params = SARIMA_AIC_df_sorted.iloc[0]

print('Best SARIMA Model Parameters:')
print(best_SARIMA_params)

SARIMA(0, 0, 0)x(0, 0, 0, 12) - AIC:4270.644636306024
SARIMA(0, 0, 0)x(0, 0, 1, 12) - AIC:3748.091530822342
SARIMA(0, 0, 0)x(0, 1, 0, 12) - AIC:2559.9293382388114
SARIMA(0, 0, 0)x(0, 1, 1, 12) - AIC:2464.4105116926403
SARIMA(0, 0, 0)x(1, 0, 0, 12) - AIC:2565.7816449072493

C:\Users\smith\anaconda3\Lib\site-packages\statsmodels\base\model.py:607: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

SARIMA(0, 0, 0)x(1, 0, 1, 12) - AIC:2558.8896289073805
SARIMA(0, 0, 0)x(1, 1, 0, 12) - AIC:2473.248093260153
SARIMA(0, 0, 0)x(1, 1, 1, 12) - AIC:2465.796423855635
SARIMA(0, 0, 1)x(0, 0, 0, 12) - AIC:3841.826405255114
SARIMA(0, 0, 1)x(0, 0, 1, 12) - AIC:3342.898760775486
SARIMA(0, 0, 1)x(0, 1, 0, 12) - AIC:2284.0782716039566
SARIMA(0, 0, 1)x(0, 1, 1, 12) - AIC:2196.977097620012
SARIMA(0, 0, 1)x(1, 0, 0, 12) - AIC:2297.274909181834
SARIMA(0, 0, 1)x(1, 0, 1, 12) - AIC:2282.3394290097667
SARIMA(0, 0, 1)x(1, 1, 0, 12) - AIC:2213.0355685823574
SARIMA(0, 0, 1)x(1, 1, 1, 12) - AIC:2196.337947812424
SARIMA(0, 1, 0)x(0, 0, 0, 12) - AIC:1844.6381562619372
SARIMA(0, 1, 0)x(0, 0, 1, 12) - AIC:1773.9419861539154
SARIMA(0, 1, 0)x(0, 1, 0, 12) - AIC:1962.4689708061962
SARIMA(0, 1, 0)x(0, 1, 1, 12) - AIC:1740.1786391888136
SARIMA(0, 1, 0)x(1, 0, 0, 12) - AIC:1781.2789674949795
SARIMA(0, 1, 0)x(1, 0, 1, 12) - AIC:1774.5650770084217
SARIMA(0, 1, 0)x(1, 1, 0, 12) - AIC:1802.5544439348114
SARIMA(0, 1, 0)x(1, 1, 1, 12) - AIC:1742.1764980255327
SARIMA(0, 1, 1)x(0, 0, 0, 12) - AIC:1841.4804755450791
SARIMA(0, 1, 1)x(0, 0, 1, 12) - AIC:1770.6699150471386
SARIMA(0, 1, 1)x(0, 1, 0, 12) - AIC:1958.750305325272
SARIMA(0, 1, 1)x(0, 1, 1, 12) - AIC:1736.916836462131
SARIMA(0, 1, 1)x(1, 0, 0, 12) - AIC:1783.118286041421
SARIMA(0, 1, 1)x(1, 0, 1, 12) - AIC:1771.7792945237495
SARIMA(0, 1, 1)x(1, 1, 0, 12) - AIC:1803.797924331097
SARIMA(0, 1, 1)x(1, 1, 1, 12) - AIC:1738.913896711998
SARIMA(1, 0, 0)x(0, 0, 0, 12) - AIC:1851.3661556378615
SARIMA(1, 0, 0)x(0, 0, 1, 12) - AIC:1783.0706069669286
SARIMA(1, 0, 0)x(0, 1, 0, 12) - AIC:1962.1201864899401
SARIMA(1, 0, 0)x(0, 1, 1, 12) - AIC:1745.6302574486635
SARIMA(1, 0, 0)x(1, 0, 0, 12) - AIC:1782.7874664876465
SARIMA(1, 0, 0)x(1, 0, 1, 12) - AIC:1784.1916095113347
SARIMA(1, 0, 0)x(1, 1, 0, 12) - AIC:1798.8642251691822
SARIMA(1, 0, 0)x(1, 1, 1, 12) - AIC:1747.6073407060012
SARIMA(1, 0, 1)x(0, 0, 0, 12) - AIC:1847.9593920802295
SARIMA(1, 0, 1)x(0, 0, 1, 12) - AIC:1777.3328169622018
SARIMA(1, 0, 1)x(0, 1, 0, 12) - AIC:1955.4376430680657
SARIMA(1, 0, 1)x(0, 1, 1, 12) - AIC:1742.084892814611
SARIMA(1, 0, 1)x(1, 0, 0, 12) - AIC:1784.635266804466
SARIMA(1, 0, 1)x(1, 0, 1, 12) - AIC:1778.540233408506
SARIMA(1, 0, 1)x(1, 1, 0, 12) - AIC:1799.4798394184065
SARIMA(1, 0, 1)x(1, 1, 1, 12) - AIC:1744.0571219614835
SARIMA(1, 1, 0)x(0, 0, 0, 12) - AIC:1846.4634498576283
SARIMA(1, 1, 0)x(0, 0, 1, 12) - AIC:1775.7865017957984
SARIMA(1, 1, 0)x(0, 1, 0, 12) - AIC:1964.3817805799167
SARIMA(1, 1, 0)x(0, 1, 1, 12) - AIC:1742.1778971242168
SARIMA(1, 1, 0)x(1, 0, 0, 12) - AIC:1775.3843712034632
SARIMA(1, 1, 0)x(1, 0, 1, 12) - AIC:1776.5256388860134
SARIMA(1, 1, 0)x(1, 1, 0, 12) - AIC:1797.9350966255488
SARIMA(1, 1, 0)x(1, 1, 1, 12) - AIC:1744.175643155097
SARIMA(1, 1, 1)x(0, 0, 0, 12) - AIC:1843.3919802036385
SARIMA(1, 1, 1)x(0, 0, 1, 12) - AIC:1772.1448903231344
SARIMA(1, 1, 1)x(0, 1, 0, 12) - AIC:1953.5902279713637
SARIMA(1, 1, 1)x(0, 1, 1, 12) - AIC:1738.7781084925
SARIMA(1, 1, 1)x(1, 0, 0, 12) - AIC:1776.9356393502662
SARIMA(1, 1, 1)x(1, 0, 1, 12) - AIC:1773.1539792805088
SARIMA(1, 1, 1)x(1, 1, 0, 12) - AIC:1797.9218761762274
SARIMA(1, 1, 1)x(1, 1, 1, 12) - AIC:1740.7768200188107
Best SARIMA Model Parameters:
param                 (0, 1, 1)
param_seasonal    (0, 1, 1, 12)
AIC                 1736.916836
Name: 27, dtype: object


# Sort the SARIMA_AIC list of dictionaries by 'AIC' in ascending order
sorted_SARIMA_AIC = sorted(SARIMA_AIC, key=lambda x: x['AIC'])

# Retrieve the parameters for the minimum AIC value
min_AIC_params = sorted_SARIMA_AIC[0]

# Print the parameters for the minimum AIC value
print("Parameters for the minimum AIC value:")
print(min_AIC_params)

Parameters for the minimum AIC value:
{'param': (0, 1, 1), 'param_seasonal': (0, 1, 1, 12), 'AIC': 1736.916836462131}


import statsmodels.api as sm

auto_SARIMA = sm.tsa.statespace.SARIMAX(train['Price'],
                                order=(0, 1, 1),
                                seasonal_order=(0, 1, 1, 12),
                                enforce_stationarity=False,
                                enforce_invertibility=False)
results_auto_SARIMA = auto_SARIMA.fit(maxiter=1000)
print(results_auto_SARIMA.summary())

                                     SARIMAX Results                                      
==========================================================================================
Dep. Variable:                              Price   No. Observations:                  311
Model:             SARIMAX(0, 1, 1)x(0, 1, 1, 12)   Log Likelihood                -865.458
Date:                            Wed, 21 Feb 2024   AIC                           1736.917
Time:                                    10:48:31   BIC                           1747.864
Sample:                                         0   HQIC                          1741.306
                                            - 311                                         
Covariance Type:                              opg                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
ma.L1          0.0360      0.039      0.924      0.356      -0.040       0.113
ma.S.L12      -0.8963      0.041    -21.956      0.000      -0.976      -0.816
sigma2        24.8874      1.329     18.727      0.000      22.283      27.492
===================================================================================
Ljung-Box (L1) (Q):                   0.00   Jarque-Bera (JB):               117.86
Prob(Q):                              0.95   Prob(JB):                         0.00
Heteroskedasticity (H):               1.96   Skew:                             0.30
Prob(H) (two-sided):                  0.00   Kurtosis:                         6.10
===================================================================================

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).


results_auto_SARIMA.plot_diagnostics();


predicted_auto_SARIMA = results_auto_SARIMA.get_forecast(steps=len(test))


predicted_auto_SARIMA.summary_frame(alpha=0.05).head()


rmse = mean_squared_error(test['Price'],predicted_auto_SARIMA.predicted_mean,squared=False)
mape = mean_absolute_percentage_error(test['Price'],predicted_auto_SARIMA.predicted_mean)
print('RMSE:',rmse,'\nMAPE:',mape)

RMSE: 19.783019045706197 
MAPE: 6.5835315802141885


temp_resultsDf = pd.DataFrame({'RMSE': rmse,'MAPE':mape}
                           ,index=['SARIMA(1,1,3)(3,0,3,6)'])


resultsDf = pd.concat([resultsDf,temp_resultsDf])

resultsDf


import pandas as pd

# Assuming 'train' is your DataFrame
train['Price'] = pd.to_numeric(train['Price'], errors='coerce')

# Drop rows with missing values
train = train.dropna()

# Compute the first discrete difference
train_diff = train['Price'].diff()

# Plot the autocorrelation and partial autocorrelation functions
plot_acf(train_diff, title='Training Data Autocorrelation', missing='drop')
plot_pacf(train_diff.dropna(), title='Training Data Partial Autocorrelation', zero=False, method='ywmle');


import statsmodels.api as sm

manual_SARIMA = sm.tsa.statespace.SARIMAX(train['Price'],
                                order=(3,1,3),
                                seasonal_order=(0, 0, 3, 6),
                                enforce_stationarity=False,
                                enforce_invertibility=False)
results_manual_SARIMA = manual_SARIMA.fit(maxiter=1000)
print(results_manual_SARIMA.summary())

                                     SARIMAX Results                                     
=========================================================================================
Dep. Variable:                             Price   No. Observations:                  311
Model:             SARIMAX(3, 1, 3)x(0, 0, 3, 6)   Log Likelihood                -854.544
Date:                           Wed, 21 Feb 2024   AIC                           1729.089
Time:                                   10:48:37   BIC                           1765.718
Sample:                                        0   HQIC                          1743.768
                                           - 311                                         
Covariance Type:                             opg                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
ar.L1          0.1989      0.919      0.216      0.829      -1.602       2.000
ar.L2         -0.5522      0.506     -1.090      0.276      -1.545       0.441
ar.L3          0.7151      0.880      0.813      0.416      -1.010       2.440
ma.L1         -0.1184      4.277     -0.028      0.978      -8.501       8.265
ma.L2          0.5525     10.608      0.052      0.958     -20.239      21.344
ma.L3         -0.7307     10.015     -0.073      0.942     -20.359      18.898
ma.S.L6    -9.865e-05      0.074     -0.001      0.999      -0.145       0.145
ma.S.L12       0.0800      0.066      1.210      0.226      -0.050       0.209
ma.S.L18       0.0523      0.071      0.738      0.460      -0.086       0.191
sigma2        21.4200    292.270      0.073      0.942    -551.418     594.258
===================================================================================
Ljung-Box (L1) (Q):                   0.34   Jarque-Bera (JB):               123.40
Prob(Q):                              0.56   Prob(JB):                         0.00
Heteroskedasticity (H):               1.82   Skew:                             0.38
Prob(H) (two-sided):                  0.00   Kurtosis:                         6.11
===================================================================================

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).


results_manual_SARIMA.plot_diagnostics()
plt.show()


predicted_manual_SARIMA = results_manual_SARIMA.get_forecast(steps=len(test))


rmse = mean_squared_error(test['Price'],predicted_manual_SARIMA.predicted_mean,squared=False)
mape = mean_absolute_percentage_error(test['Price'],predicted_manual_SARIMA.predicted_mean)
print('RMSE:',rmse,'\nMAPE:',mape)

RMSE: 26.326591909036644 
MAPE: 8.72893161271512


temp_resultsDf = pd.DataFrame({'RMSE': [rmse],'MAPE':mape}
                           ,index=['SARIMA(3,1,3)(0,0,3,6)'])


resultsDf = pd.concat([resultsDf,temp_resultsDf])

resultsDf


import pandas as pd

# Assuming `df` is your DataFrame
df['Price'] = pd.to_numeric(df['Price'], errors='coerce')

# Remove NaN values
df = df.dropna()

# Now, create and fit your SARIMAX model
full_data_model = sm.tsa.statespace.SARIMAX(df['Price'],
                                order=(1,1,3),
                                seasonal_order=(3, 0, 3, 6),
                                enforce_stationarity=False,
                                enforce_invertibility=False)
results_full_data_model = full_data_model.fit(maxiter=1000)
print(results_full_data_model.summary())

                                     SARIMAX Results                                     
=========================================================================================
Dep. Variable:                             Price   No. Observations:                  311
Model:             SARIMAX(1, 1, 3)x(3, 0, 3, 6)   Log Likelihood                -853.129
Date:                           Wed, 21 Feb 2024   AIC                           1728.257
Time:                                   10:48:42   BIC                           1768.550
Sample:                                        0   HQIC                          1744.404
                                           - 311                                         
Covariance Type:                             opg                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
ar.L1         -0.4828      0.245     -1.974      0.048      -0.962      -0.003
ma.L1          0.5639      0.230      2.451      0.014       0.113       1.015
ma.L2          0.0155      0.048      0.323      0.746      -0.078       0.109
ma.L3         -0.1416      0.050     -2.860      0.004      -0.239      -0.045
ar.S.L6        0.7189      0.394      1.826      0.068      -0.053       1.490
ar.S.L12       0.1513      0.493      0.307      0.759      -0.816       1.118
ar.S.L18      -0.4221      0.255     -1.656      0.098      -0.922       0.077
ma.S.L6       -0.8650      0.397     -2.179      0.029      -1.643      -0.087
ma.S.L12       0.0136      0.556      0.024      0.980      -1.075       1.103
ma.S.L18       0.5366      0.340      1.580      0.114      -0.129       1.202
sigma2        20.6344      1.435     14.379      0.000      17.822      23.447
===================================================================================
Ljung-Box (L1) (Q):                   0.09   Jarque-Bera (JB):               100.28
Prob(Q):                              0.77   Prob(JB):                         0.00
Heteroskedasticity (H):               0.57   Skew:                            -0.20
Prob(H) (two-sided):                  0.01   Kurtosis:                         5.86
===================================================================================

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).


predicted_manual_SARIMA_full_data = results_full_data_model.get_forecast(steps=20)


pred_full_manual_SARIMA_date = predicted_manual_SARIMA_full_data.summary_frame(alpha=0.05)
pred_full_manual_SARIMA_date.head()


rmse = mean_squared_error(df['Price'],results_full_data_model.fittedvalues,squared=False)
print('RMSE of the Full Model',rmse)

RMSE of the Full Model 16.587446353653686


axis = df['Price'].plot(label='Observed')
pred_full_manual_SARIMA_date['mean'].plot(ax=axis, label='Forecast', alpha=0.7)

axis.set_xlabel('Year-Months')
axis.set_ylabel('Sales_quantity')
plt.legend(loc='best')
plt.grid();


resultsDf


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from statsmodels.tsa.arima.model import ARIMA
from statsmodels.tsa.statespace.sarimax import SARIMAX
from sklearn.metrics import mean_squared_error
from math import sqrt

# Generate sample data
train_final = pd.DataFrame({
    'Price': np.random.rand(50),
    'Date': pd.date_range(start='2021-03-31', periods=50, freq='D')
})

test_final = pd.DataFrame({
    'Price': np.random.rand(10),
    'Date': pd.date_range(start='2021-03-31', periods=10, freq='D')
})

# Define the orders and seasonal orders as tuples
p = 1  # Number of AR terms
d = 1  # Number of differences (integration)
q = 1  # Number of MA terms
seasonal_p = 1  # Number of seasonal AR terms
seasonal_d = 0  # Number of seasonal differences (integration)
seasonal_q = 1  # Number of seasonal MA terms
s = 12  # Seasonal period

# Define and fit ARIMA model
arima_model = ARIMA(train_final['Price'], order=(p, d, q), seasonal_order=(seasonal_p, seasonal_d, seasonal_q, s))
arima_fit = arima_model.fit()

# Generate predictions
pred_ARIMA = arima_fit.predict(start=len(train_final), end=len(train_final) + len(test_final) - 1)

# Define and fit SARIMA model
sarima_model = SARIMAX(train_final['Price'], order=(p, d, q), seasonal_order=(seasonal_p, seasonal_d, seasonal_q, s))
sarima_fit = sarima_model.fit()

# Generate predictions
pred_SARIMA = sarima_fit.predict(start=len(train_final), end=len(train_final) + len(test_final) - 1)

# Define and fit Prophet model

# Calculate RMSE
rmse_arima = sqrt(mean_squared_error(test_final['Price'], pred_ARIMA))
rmse_sarima = sqrt(mean_squared_error(test_final['Price'], pred_SARIMA))

# Plotting
plt.plot(train_final['Date'], train_final['Price'], label='Training Data')
plt.plot(test_final['Date'], test_final['Price'], label='Test Data')
plt.plot(test_final['Date'], pred_ARIMA, label='Predicted Data - ARIMA')
plt.plot(test_final['Date'], pred_SARIMA, label='Predicted Data - SARIMA')
plt.legend(loc='best')
plt.grid()
plt.show()

# Print RMSE
print("Root Mean Squared Error (RMSE) of the ARIMA model:", rmse_arima)
print("Root Mean Squared Error (RMSE) of the SARIMA model:", rmse_sarima)

Root Mean Squared Error (RMSE) of the ARIMA model: 0.3177718764476827
Root Mean Squared Error (RMSE) of the SARIMA model: 0.3177718764476827


df_prophet = pd.DataFrame()
df_prophet['ds'] = pd.to_datetime(df['Date'])
df_prophet['y'] = df['Price']
df_prophet['High']= df['High']
df_prophet['Low']=df['Low']
df_prophet.head()


import pandas as pd
from prophet import Prophet
from prophet.plot import plot_plotly, plot_components_plotly
import plotly.graph_objs as go

# Drop rows with NaN values
df_prophet_filtered = df_prophet.dropna(subset=['y'])

# Assuming df_prophet is your DataFrame with 'ds' and 'y' columns
prophet = Prophet(seasonality_mode='additive', weekly_seasonality=True)
prophet.add_country_holidays(country_name='US')

# Filter the data if needed (assuming df_prophet is already defined)
df_prophet_filtered = df_prophet[df_prophet['ds'] <= pd.to_datetime('2021-03-31')]

# Fit the model
prophet.fit(df_prophet_filtered)

# Create future DataFrame for forecasting
future = prophet.make_future_dataframe(periods=43, freq='B')

# Make predictions
forecast = prophet.predict(future)

# Plot the forecast using plotly
fig = plot_plotly(prophet, forecast)
fig.show()

# Plot the forecast components using plotly
fig_components = plot_components_plotly(prophet, forecast)
fig_components.show()

10:48:43 - cmdstanpy - INFO - Chain [1] start processing
10:48:44 - cmdstanpy - INFO - Chain [1] done processing


import holidays
for date, name in sorted(holidays.US(state='NY', years=2021).items()):
   print(date, name)

2021-01-01 New Year's Day
2021-01-18 Martin Luther King Jr. Day
2021-02-12 Lincoln's Birthday
2021-02-15 Susan B. Anthony Day; Washington's Birthday
2021-05-31 Memorial Day
2021-06-18 Juneteenth National Independence Day (observed)
2021-06-19 Juneteenth National Independence Day
2021-07-04 Independence Day
2021-07-05 Independence Day (observed)
2021-09-06 Labor Day
2021-10-11 Columbus Day
2021-11-02 Election Day
2021-11-11 Veterans Day
2021-11-25 Thanksgiving
2021-12-24 Christmas Day (observed)
2021-12-25 Christmas Day
2021-12-31 New Year's Day (observed)


forecast


import numpy as np
from sklearn.metrics import mean_squared_error
from math import sqrt

# Assuming 'test_final' and 'forecast' are properly defined DataFrames
test_price_values = test_final['Price'].values
forecast_trimmed = forecast['yhat'].values[-len(test_price_values):]  # Trims 'forecast' to match the length of 'test_final'

rmse = sqrt(mean_squared_error(test_price_values, forecast_trimmed))
print(rmse)

294.07462064541266


resultsDf_temp = pd.DataFrame({'Test RMSE': rmse}
                           ,index=['Prophet'])

resultsDf = pd.concat([resultsDf, resultsDf_temp])
resultsDf


from prophet import Prophet

# Convert the 'Date' column to datetime
train_final['Date'] = pd.to_datetime(train_final['Date'])
test_final['Date'] = pd.to_datetime(test_final['Date'])

# Rename the 'Date' column to 'ds' and the 'Price' column to 'y' for Prophet
train_final = train_final.rename(columns={'Date': 'ds', 'Price': 'y'})
test_final = test_final.rename(columns={'Date': 'ds', 'Price': 'y'})

# Define and fit Prophet model
prophet_model = Prophet()
prophet_model.fit(train_final)

# Generate predictions
future = pd.DataFrame({'ds': test_final['ds']})
pred_prophet = prophet_model.predict(future)

# Plotting
plt.plot(train_final['ds'], train_final['y'], label='Training Data')
plt.plot(test_final['ds'], test_final['y'], label='Test Data')
plt.plot(test_final['ds'], pred_ARIMA, label='Predicted Data - ARIMA')
plt.plot(test_final['ds'], pred_SARIMA, label='Predicted Data - SARIMA')
plt.plot(test_final['ds'], pred_prophet['yhat'], label='Predicted Data - Prophet')
plt.legend(loc='best')
plt.grid()
plt.show()

# Calculate RMSE
rmse_prophet = sqrt(mean_squared_error(test_final['y'], pred_prophet['yhat']))

# Print RMSE
print("Root Mean Squared Error (RMSE) of the ARIMA model:", rmse_arima)
print("Root Mean Squared Error (RMSE) of the SARIMA model:", rmse_sarima)
print("Root Mean Squared Error (RMSE) of the Prophet model:", rmse_prophet)

10:48:45 - cmdstanpy - INFO - Chain [1] start processing
10:48:45 - cmdstanpy - INFO - Chain [1] done processing

Root Mean Squared Error (RMSE) of the ARIMA model: 0.3177718764476827
Root Mean Squared Error (RMSE) of the SARIMA model: 0.3177718764476827
Root Mean Squared Error (RMSE) of the Prophet model: 0.35281333708257934


future = df_prophet[['ds']].copy()  # Make sure you only have the 'ds' column
future = prophet.make_future_dataframe(periods=len(df_prophet), freq=us_bd)
future['High'] = df_prophet['High']
future['Low'] = df_prophet['Low']

print(f"Number of missing values in 'High' before filling: {future['High'].isnull().sum()}")
print(f"Number of missing values in 'Low' before filling: {future['Low'].isnull().sum()}")

# Fill missing values with 'ffill' method
future['High'] = future['High'].fillna(method='ffill')
future['Low'] = future['Low'].fillna(method='ffill')

print(f"Number of missing values in 'High' after filling: {future['High'].isnull().sum()}")
print(f"Number of missing values in 'Low' after filling: {future['Low'].isnull().sum()}")

# Fit Prophet model as usual
prophet = Prophet()
prophet.add_country_holidays(country_name='US')
prophet.add_regressor('High')
prophet.add_regressor('Low')
prophet.fit(df_prophet[df_prophet['ds'] < pd.to_datetime('2021-03-31')])

# Predict and plot
forecast = prophet.predict(future)
fig = prophet.plot(forecast)
a = add_changepoints_to_plot(fig.gca(), prophet, forecast)
plt.show()
fig2 = prophet.plot_components(forecast)
plt.show()

10:48:45 - cmdstanpy - INFO - Chain [1] start processing

Number of missing values in 'High' before filling: 311
Number of missing values in 'Low' before filling: 311
Number of missing values in 'High' after filling: 0
Number of missing values in 'Low' after filling: 0

10:48:45 - cmdstanpy - INFO - Chain [1] done processing


print(test_final.columns)

Index(['y', 'ds'], dtype='object')


rmse = sqrt(mean_squared_error(test_final['y'],forecast.tail(10)['yhat']))
print(rmse)

291.58876330022275


resultsDf_temp = pd.DataFrame({'Test RMSE': rmse}
                           ,index=['Prophet - with exog variables_2.0'])

resultsDf = pd.concat([resultsDf, resultsDf_temp])
resultsDf


import pandas as pd
from prophet import Prophet

# Rename the 'Date' column to 'ds' and the 'Price' column to 'y' for Prophet
train_final = train_final.rename(columns={'Date': 'ds', 'Price': 'y'})
test_final = test_final.rename(columns={'Date': 'ds', 'Price': 'y'})

# Convert the 'Date' column to datetime
train_final['ds'] = pd.to_datetime(train_final['ds'])
test_final['ds'] = pd.to_datetime(test_final['ds'])

# Define and fit Prophet model
prophet_model = Prophet()
prophet_model.fit(train_final)

# Generate predictions
future = pd.DataFrame({'ds': test_final['ds']})
pred_prophet = prophet_model.predict(future)

# Generate predictions without exogenous variables
future_no_exog = pd.DataFrame({'ds': test_final['ds']})
pred_prophet_no_exog = prophet_model.predict(future_no_exog)

# Plotting
plt.plot(train_final['ds'], train_final['y'], label='Training Data')
plt.plot(test_final['ds'], test_final['y'], label='Test Data')
plt.plot(test_final['ds'], pred_ARIMA, label='Predicted Data - ARIMA')
plt.plot(test_final['ds'], pred_SARIMA, label='Predicted Data - SARIMA')

# Plotting Prophet predictions
plt.plot(pred_prophet['ds'], pred_prophet['yhat'], label='Predicted Data - Prophet')
plt.plot(pred_prophet_no_exog['ds'], pred_prophet_no_exog['yhat'], label='Predicted Data - Prophet - without exog variables')

plt.legend(loc='best')
plt.grid()
plt.show()

# Calculate RMSE for both cases
rmse_prophet = sqrt(mean_squared_error(test_final['y'], pred_prophet['yhat']))
rmse_prophet_no_exog = sqrt(mean_squared_error(test_final['y'], pred_prophet_no_exog['yhat']))

# Print RMSE
print("Root Mean Squared Error (RMSE) of the ARIMA model:", rmse_arima)
print("Root Mean Squared Error (RMSE) of the SARIMA model:", rmse_sarima)
print("Root Mean Squared Error (RMSE) of the Prophet model with exog variables:", rmse_prophet)
print("Root Mean Squared Error (RMSE) of the Prophet model without exog variables:", rmse_prophet_no_exog)
print("Difference in RMSE between Prophet models (with and without exog variables):", rmse_prophet - rmse_prophet_no_exog)

10:48:47 - cmdstanpy - INFO - Chain [1] start processing
10:48:47 - cmdstanpy - INFO - Chain [1] done processing

Root Mean Squared Error (RMSE) of the ARIMA model: 0.3177718764476827
Root Mean Squared Error (RMSE) of the SARIMA model: 0.3177718764476827
Root Mean Squared Error (RMSE) of the Prophet model with exog variables: 0.35281333708257934
Root Mean Squared Error (RMSE) of the Prophet model without exog variables: 0.35281333708257934
Difference in RMSE between Prophet models (with and without exog variables): 0.0


future = df_prophet[['ds']].copy()  # Make sure you only have the 'ds' column
future = prophet.make_future_dataframe(periods=len(df_prophet), freq=us_bd)
future['High'] = df_prophet['High']
future['Low'] = df_prophet['Low']

print(f"Number of missing values in 'High' before filling: {future['High'].isnull().sum()}")
print(f"Number of missing values in 'Low' before filling: {future['Low'].isnull().sum()}")

# Fill missing values with 'ffill' method
future['High'] = future['High'].fillna(method='ffill')
future['Low'] = future['Low'].fillna(method='ffill')

print(f"Number of missing values in 'High' after filling: {future['High'].isnull().sum()}")
print(f"Number of missing values in 'Low' after filling: {future['Low'].isnull().sum()}")

# Fit Prophet model as usual
prophet = Prophet()
prophet.add_country_holidays(country_name='US')
prophet.add_regressor('High')
prophet.add_regressor('Low')
prophet.fit(df_prophet[df_prophet['ds'] < pd.to_datetime('2021-03-31')])

# Predict and plot
forecast = prophet.predict(future)
fig = prophet.plot(forecast)
a = add_changepoints_to_plot(fig.gca(), prophet, forecast)
plt.show()
fig2 = prophet.plot_components(forecast)
plt.show()

10:48:48 - cmdstanpy - INFO - Chain [1] start processing

Number of missing values in 'High' before filling: 310
Number of missing values in 'Low' before filling: 310
Number of missing values in 'High' after filling: 0
Number of missing values in 'Low' after filling: 0

10:48:48 - cmdstanpy - INFO - Chain [1] done processing


import pandas as pd

# Define a range of dates using the pd.date_range function
weekend = pd.DataFrame({
  'holiday': 'Weekend',
  'ds': pd.date_range(start='2020-01-05', end='2021-03-28', freq='W'),
  'lower_window': -1,
  'upper_window': 0,
})

print(weekend)

    holiday         ds  lower_window  upper_window
0   Weekend 2020-01-05            -1             0
1   Weekend 2020-01-12            -1             0
2   Weekend 2020-01-19            -1             0
3   Weekend 2020-01-26            -1             0
4   Weekend 2020-02-02            -1             0
..      ...        ...           ...           ...
60  Weekend 2021-02-28            -1             0
61  Weekend 2021-03-07            -1             0
62  Weekend 2021-03-14            -1             0
63  Weekend 2021-03-21            -1             0
64  Weekend 2021-03-28            -1             0

[65 rows x 4 columns]


#Cross Validation of developed time series model. This is a function is for cross validation. Its creates train and test dataset. Initial days define train data
#and horizon data define test data.
from prophet.diagnostics import cross_validation
df_cv = cross_validation(prophet, initial='208 days',period = '25 days', horizon = '43 days')

from prophet.diagnostics import performance_metrics
df_p = performance_metrics(df_cv)

  0%|          | 0/9 [00:00<?, ?it/s]

10:48:49 - cmdstanpy - INFO - Chain [1] start processing
10:48:49 - cmdstanpy - INFO - Chain [1] done processing
10:48:50 - cmdstanpy - INFO - Chain [1] start processing
10:48:50 - cmdstanpy - INFO - Chain [1] done processing
10:48:50 - cmdstanpy - INFO - Chain [1] start processing
10:48:50 - cmdstanpy - INFO - Chain [1] done processing
10:48:50 - cmdstanpy - INFO - Chain [1] start processing
10:48:50 - cmdstanpy - INFO - Chain [1] done processing
10:48:50 - cmdstanpy - INFO - Chain [1] start processing
10:48:50 - cmdstanpy - INFO - Chain [1] done processing
10:48:51 - cmdstanpy - INFO - Chain [1] start processing
10:48:51 - cmdstanpy - INFO - Chain [1] done processing
10:48:51 - cmdstanpy - INFO - Chain [1] start processing
10:48:51 - cmdstanpy - INFO - Chain [1] done processing
10:48:51 - cmdstanpy - INFO - Chain [1] start processing
10:48:51 - cmdstanpy - INFO - Chain [1] done processing
10:48:51 - cmdstanpy - INFO - Chain [1] start processing
10:48:51 - cmdstanpy - INFO - Chain [1] done processing


# Time series model performance (Cross validation performance metrics can be visualized with plot_cross_validation_metric,
# here shown for MAPE. Dots show the absolute percent error for each prediction in df_cv. The blue line shows the MAPE,
#where the mean is taken over a rolling window of the dots.)
from prophet.plot import plot_cross_validation_metric
fig = plot_cross_validation_metric(df_cv, metric='mape')


df_cv


df_pd = df_date.copy()
print(df_pd)

             Date   Price Time_Stamp
0    Jan 03, 2020  255.00 2020-01-03
1    Jan 06, 2020  253.90 2020-01-06
2    Jan 08, 2020  259.15 2020-01-07
3    Jan 09, 2020  257.99 2020-01-08
4    Jan 10, 2020  258.19 2020-01-09
..            ...     ...        ...
306  Mar 25, 2021  286.66 2021-03-24
307  Mar 26, 2021  291.22 2021-03-25
308  Mar 29, 2021  293.30 2021-03-26
309  Mar 30, 2021  293.25 2021-03-29
310  Mar 31, 2021  291.02 2021-03-30

[311 rows x 3 columns]


# We have considered Price as a dependent variable for this session. Readers are advised to look at  'Vol.','open' price as dependent variable for univariate Time Series Prediction
df_date = df[['Date','Price']]


df_pd['Date'] = pd.to_datetime(df_pd['Date'])


df_pd['Price'] = df_pd['Price'].astype(float)


df_pd.dtypes

Date          datetime64[ns]
Price                float64
Time_Stamp    datetime64[ns]
dtype: object


import pandas as pd

# Assuming df['Date'] contains strings in the format 'YYYY-MM-DD'
df['Date'] = pd.to_datetime(df['Date'])

specific_date_str = "2020-12-31"
specific_date = pd.to_datetime(specific_date_str)

filtered_df = df[df['Date'] <= specific_date]


filtered_df.rename(columns={'Date': 'ds', 'Price': 'y'}, inplace=True)


# Convert the 'date' column to datetime objects
filtered_df['ds'] = pd.to_datetime(filtered_df['ds'])


print(filtered_df)

            ds       y    Open    High     Low Time_Stamp
0   2020-01-03  255.00  255.99  258.19   253.7 2020-01-03
1   2020-01-06  253.90  254.75  254.84   251.4 2020-01-06
2   2020-01-08  259.15  253.57  259.15  253.03 2020-01-07
3   2020-01-09  257.99   259.4  261.76  257.01 2020-01-08
4   2020-01-10  258.19  257.86  259.25  256.88 2020-01-09
..         ...     ...     ...     ...     ...        ...
245 2020-12-24  269.32  271.98  272.82  267.13 2020-12-23
246 2020-12-25  270.64  269.97     271  268.28 2020-12-24
247 2020-12-28  274.70   270.9  275.19  269.73 2020-12-28
248 2020-12-29  272.00  275.91  276.43  271.26 2020-12-29
249 2020-12-30  271.65  272.49  274.24  270.12 2020-12-30

[250 rows x 6 columns]


df_Test = df[df['Date'] > specific_date]


df_Test


# plot the time series
plt.plot(filtered_df['ds'],filtered_df['y'])
# plt.plot(df_sorted['ds'])
plt.xlabel('date')
plt.ylabel('price')
plt.title('Current Prices')
plt.show()


train = df_pd[df_pd['Date'] < specific_date]


print(train.shape)

(250, 3)


test = df_pd[df_pd['Date'] >= specific_date]


print(test.shape)

(61, 3)


import pandas as pd

# Assuming df is your DataFrame and specific_date_str is the specific date string
specific_date_str = "2020-12-31"
specific_date = pd.to_datetime(specific_date_str)

# Filter the DataFrame before the Q1 of the second year
filtered_df = df[df['Date'] <= specific_date]


filtered_df.rename(columns={'Date': 'ds', 'Price': 'y'}, inplace=True)


# Convert the 'date' column to datetime objects filtered_df['ds'] = pd.to_datetime(filtered_df['ds'])


print(filtered_df)

            ds       y    Open    High     Low Time_Stamp
0   2020-01-03  255.00  255.99  258.19   253.7 2020-01-03
1   2020-01-06  253.90  254.75  254.84   251.4 2020-01-06
2   2020-01-08  259.15  253.57  259.15  253.03 2020-01-07
3   2020-01-09  257.99   259.4  261.76  257.01 2020-01-08
4   2020-01-10  258.19  257.86  259.25  256.88 2020-01-09
..         ...     ...     ...     ...     ...        ...
245 2020-12-24  269.32  271.98  272.82  267.13 2020-12-23
246 2020-12-25  270.64  269.97     271  268.28 2020-12-24
247 2020-12-28  274.70   270.9  275.19  269.73 2020-12-28
248 2020-12-29  272.00  275.91  276.43  271.26 2020-12-29
249 2020-12-30  271.65  272.49  274.24  270.12 2020-12-30

[250 rows x 6 columns]


df_Test = df[df['Date'] > specific_date]


print(test.shape)

(61, 3)


from sklearn.preprocessing import MinMaxScaler

# Initialize the MinMaxScaler
scaler = MinMaxScaler(feature_range=(0, 1))

# Reshape the 'Price' values and scale them
df_scaled = scaler.fit_transform(df_pd['Price'].values.reshape(-1,1))

# Create a new DataFrame with the scaled values
df_scaled = pd.DataFrame(df_scaled, columns=['Price'])

# Include the 'Date' column from the original DataFrame
df_scaled['Date'] = df_pd['Date']

# Display the scaled DataFrame
print(df_scaled)

        Price       Date
0    0.678232 2020-01-03
1    0.668991 2020-01-06
2    0.713098 2020-01-08
3    0.703352 2020-01-09
4    0.705032 2020-01-10
..        ...        ...
306  0.944216 2021-03-25
307  0.982525 2021-03-26
308  1.000000 2021-03-29
309  0.999580 2021-03-30
310  0.980845 2021-03-31

[311 rows x 2 columns]


from sklearn.preprocessing import MinMaxScaler

# Initialize the MinMaxScaler
scaler = MinMaxScaler(feature_range=(0, 1))

# Scale the 'Price' values
df_scaled = scaler.fit_transform(df_pd['Price'].values.reshape(-1,1))

# Note: You have commented out the following lines, but I've left them here for reference.
# df_scaled_test = scaler.fit_transform(test['Price'].values.reshape(-1,1))
# df_scaled = pd.DataFrame(df_scaled, columns=['Price'])
# df_scaled['Date'] = df['Date']


# Define the fraction of data for training
train_fraction = 0.7  # Assuming 70% of the data is used for training

# Calculate the size of the training set
train_size = int(len(df_scaled) * train_fraction)

# Split the data into training and testing sets
train_data, test_data = df_scaled[0:train_size], df_scaled[train_size:]

# Function to create sequences of data for training
def create_dataset(dataset, look_back=1):
    X, Y = [], []
    for i in range(len(dataset) - look_back):
        X.append(dataset[i:(i + look_back), 0])
        Y.append(dataset[i + look_back, 0])
    return np.array(X), np.array(Y)

# Number of previous time steps to use for prediction
look_back = 30

# Create sequences of data for training
X_train, y_train = create_dataset(train_data, look_back)
X_test, y_test = create_dataset(test_data, look_back)


# Reshape data for LSTM input
X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 1))
X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))


X_train.shape, y_train.shape, X_test.shape,y_test.shape

((187, 30, 1), (187,), (64, 30, 1), (64,))


import numpy as np

from keras.models import Sequential
from keras.layers import LSTM, Dense

# Generate synthetic data
n_samples = 1000
n_features = 10
X_train = np.random.rand(n_samples, n_features)
y_train = np.random.rand(n_samples)

# Create and compile the LSTM model
model = Sequential()
model.add(LSTM(50, input_shape=(n_features, 1)))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam')

# Train the model
model.fit(X_train, y_train, epochs=100, batch_size=16)

WARNING:tensorflow:From C:\Users\smith\anaconda3\Lib\site-packages\keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.

WARNING:tensorflow:From C:\Users\smith\anaconda3\Lib\site-packages\keras\src\backend.py:873: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.

WARNING:tensorflow:From C:\Users\smith\anaconda3\Lib\site-packages\keras\src\optimizers\__init__.py:309: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.

Epoch 1/100
WARNING:tensorflow:From C:\Users\smith\anaconda3\Lib\site-packages\keras\src\utils\tf_utils.py:492: The name tf.ragged.RaggedTensorValue is deprecated. Please use tf.compat.v1.ragged.RaggedTensorValue instead.

63/63 [==============================] - 3s 4ms/step - loss: 0.0993
Epoch 2/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0888
Epoch 3/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0874
Epoch 4/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0877
Epoch 5/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0869
Epoch 6/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0871
Epoch 7/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0861
Epoch 8/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0862
Epoch 9/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0859
Epoch 10/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0877
Epoch 11/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0860
Epoch 12/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0864
Epoch 13/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0861
Epoch 14/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0872
Epoch 15/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0868
Epoch 16/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0862
Epoch 17/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 18/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0876
Epoch 19/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0863
Epoch 20/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 21/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0863
Epoch 22/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 23/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 24/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 25/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0862
Epoch 26/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0866
Epoch 27/100
63/63 [==============================] - 0s 4ms/step - loss: 0.0858
Epoch 28/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0863
Epoch 29/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0860
Epoch 30/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 31/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 32/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 33/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 34/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0854
Epoch 35/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0865
Epoch 36/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 37/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 38/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 39/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0859
Epoch 40/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0867
Epoch 41/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0864
Epoch 42/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0860
Epoch 43/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 44/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0868
Epoch 45/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 46/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0862
Epoch 47/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 48/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 49/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0853
Epoch 50/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 51/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0859
Epoch 52/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 53/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0867
Epoch 54/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 55/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 56/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 57/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0854
Epoch 58/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 59/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0862
Epoch 60/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 61/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0859
Epoch 62/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 63/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0882
Epoch 64/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 65/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 66/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 67/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 68/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 69/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 70/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 71/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 72/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 73/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 74/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0854
Epoch 75/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0853
Epoch 76/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0860
Epoch 77/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 78/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 79/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 80/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0859
Epoch 81/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0860
Epoch 82/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 83/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0855
Epoch 84/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 85/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0853
Epoch 86/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0861
Epoch 87/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 88/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0860
Epoch 89/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0858
Epoch 90/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 91/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 92/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 93/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0857
Epoch 94/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0865
Epoch 95/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0852
Epoch 96/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0859
Epoch 97/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 98/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0860
Epoch 99/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0856
Epoch 100/100
63/63 [==============================] - 0s 5ms/step - loss: 0.0852

<keras.src.callbacks.History at 0x1bb00e0d2d0>


import numpy as np

# Assuming X_train is your training dataset and has shape (num_samples, num_timesteps, num_features)
# Add a new axis for the number of samples
X_train_reshaped = np.expand_dims(X_train[-1], axis=0)

# Make predictions on the reshaped training data
predict_X = model.predict(X_train_reshaped)

1/1 [==============================] - 1s 623ms/step


from sklearn.preprocessing import MinMaxScaler

# Assuming scaler is the MinMaxScaler object you used to scale your training data
# Make predictions and inverse transform
predict_X = model.predict(X_train)
predict_X = scaler.inverse_transform(predict_X)

32/32 [==============================] - 0s 3ms/step


print(predict_X)

[[241.89792]
 [240.29874]
 [238.24861]
 [239.1619 ]
 [237.73277]
 [239.45108]
 [238.97891]
 [239.9829 ]
 [241.88364]
 [240.58308]
 [241.31998]
 [241.08449]
 [237.59492]
 [240.55003]
 [241.52286]
 [240.78526]
 [239.76938]
 [240.89325]
 [239.21474]
 [241.24654]
 [237.14569]
 [239.14972]
 [240.41707]
 [238.45375]
 [240.25597]
 [239.64676]
 [240.68503]
 [238.50958]
 [237.56644]
 [241.36626]
 [241.31192]
 [241.0866 ]
 [238.54077]
 [238.16055]
 [240.74661]
 [241.48143]
 [239.57904]
 [239.24377]
 [239.07532]
 [241.62224]
 [237.56242]
 [238.98027]
 [241.71013]
 [239.36856]
 [240.26138]
 [241.91446]
 [239.72696]
 [239.84853]
 [242.31458]
 [241.07549]
 [240.81702]
 [239.94807]
 [240.11574]
 [240.70398]
 [239.6833 ]
 [242.3335 ]
 [241.87584]
 [242.2829 ]
 [239.56148]
 [235.48973]
 [237.53725]
 [242.05846]
 [242.30847]
 [240.29393]
 [241.74387]
 [239.32051]
 [238.38463]
 [242.5669 ]
 [241.1704 ]
 [237.41339]
 [241.78459]
 [240.25693]
 [237.06288]
 [240.14883]
 [240.61473]
 [240.65932]
 [241.94753]
 [241.77567]
 [240.89609]
 [242.15236]
 [239.80127]
 [239.64839]
 [240.71698]
 [241.36096]
 [241.78162]
 [242.32811]
 [241.10565]
 [240.64519]
 [241.23958]
 [241.38649]
 [239.60217]
 [241.9652 ]
 [239.53726]
 [238.20496]
 [242.28296]
 [240.52386]
 [239.29366]
 [238.26027]
 [238.31998]
 [240.74982]
 [241.00613]
 [242.29057]
 [242.00513]
 [238.96518]
 [238.97334]
 [243.02005]
 [241.16772]
 [240.9322 ]
 [239.94916]
 [239.5734 ]
 [238.48067]
 [238.16267]
 [241.14369]
 [240.81458]
 [238.56972]
 [237.11594]
 [241.46959]
 [238.48721]
 [241.73962]
 [240.06682]
 [241.02663]
 [242.11664]
 [241.47714]
 [240.67107]
 [239.2115 ]
 [236.77013]
 [238.96538]
 [239.54518]
 [241.07835]
 [242.69128]
 [240.88533]
 [239.18622]
 [240.21677]
 [240.14966]
 [240.70833]
 [239.5938 ]
 [239.64954]
 [238.6851 ]
 [240.00677]
 [239.82225]
 [240.60905]
 [239.5863 ]
 [238.3743 ]
 [241.16782]
 [238.21251]
 [237.29799]
 [241.71527]
 [242.05103]
 [238.10124]
 [240.0789 ]
 [238.84772]
 [240.28871]
 [241.91617]
 [241.04092]
 [241.33028]
 [242.98616]
 [240.30435]
 [238.25426]
 [238.8133 ]
 [241.88661]
 [241.30928]
 [239.97275]
 [238.98158]
 [239.61633]
 [240.95035]
 [236.82562]
 [240.96521]
 [238.70653]
 [244.14453]
 [240.24916]
 [240.27063]
 [241.27614]
 [241.16403]
 [238.67963]
 [238.71907]
 [240.04648]
 [241.98604]
 [239.50197]
 [240.17647]
 [238.68695]
 [243.79218]
 [239.91678]
 [241.32014]
 [240.64618]
 [241.54732]
 [241.3758 ]
 [240.64946]
 [238.686  ]
 [240.95293]
 [241.16293]
 [241.13069]
 [240.05663]
 [241.74782]
 [237.61469]
 [240.81616]
 [241.01465]
 [242.9531 ]
 [239.34673]
 [242.71295]
 [241.7609 ]
 [239.05281]
 [238.60927]
 [239.18968]
 [239.63159]
 [239.65489]
 [241.45453]
 [236.82034]
 [240.18825]
 [240.87164]
 [240.29169]
 [241.21318]
 [238.61415]
 [243.56598]
 [241.25974]
 [238.88301]
 [239.69153]
 [237.626  ]
 [238.98459]
 [243.11072]
 [240.83742]
 [240.60071]
 [241.7294 ]
 [241.54979]
 [241.94412]
 [238.32619]
 [240.5282 ]
 [242.65503]
 [239.10252]
 [239.25124]
 [241.69974]
 [240.5778 ]
 [241.63129]
 [240.02933]
 [240.2392 ]
 [241.34705]
 [238.35284]
 [241.59451]
 [239.25844]
 [240.97432]
 [240.7382 ]
 [240.44604]
 [238.88254]
 [240.07672]
 [242.38303]
 [239.23404]
 [239.88043]
 [238.439  ]
 [238.97139]
 [240.73347]
 [241.68758]
 [240.2105 ]
 [236.1832 ]
 [243.24123]
 [239.04013]
 [241.2601 ]
 [241.00418]
 [241.99532]
 [242.21223]
 [236.18507]
 [241.17737]
 [242.154  ]
 [239.74518]
 [242.28424]
 [239.25015]
 [239.64613]
 [238.16501]
 [239.99246]
 [242.55281]
 [241.18317]
 [238.51596]
 [241.6732 ]
 [238.96759]
 [240.45303]
 [238.41367]
 [242.0001 ]
 [240.84193]
 [239.11627]
 [241.14317]
 [241.24792]
 [240.1475 ]
 [241.54256]
 [240.01653]
 [240.21617]
 [239.45897]
 [241.02917]
 [240.53569]
 [238.06154]
 [237.2153 ]
 [239.98982]
 [239.71123]
 [238.55745]
 [237.7567 ]
 [237.9975 ]
 [238.08966]
 [238.85323]
 [240.80501]
 [240.7715 ]
 [240.65024]
 [239.58752]
 [242.44293]
 [238.01709]
 [237.63495]
 [240.41092]
 [239.5289 ]
 [239.66838]
 [239.58736]
 [241.92737]
 [240.26703]
 [238.30878]
 [239.32085]
 [241.40239]
 [240.80135]
 [241.79527]
 [238.91316]
 [239.61719]
 [238.78584]
 [241.44655]
 [237.23674]
 [240.49509]
 [238.09016]
 [239.01057]
 [239.83578]
 [238.24458]
 [238.90291]
 [240.5385 ]
 [239.85616]
 [240.20514]
 [241.61115]
 [240.52974]
 [240.06557]
 [242.35686]
 [239.39006]
 [238.80347]
 [240.7783 ]
 [238.74953]
 [241.19894]
 [242.71881]
 [241.27094]
 [240.5847 ]
 [241.6238 ]
 [241.97287]
 [240.54807]
 [238.15933]
 [239.56917]
 [241.26506]
 [241.15251]
 [238.4079 ]
 [239.7503 ]
 [239.01462]
 [238.60728]
 [239.8034 ]
 [241.64508]
 [242.30424]
 [239.2246 ]
 [239.7503 ]
 [243.44714]
 [241.78343]
 [239.91931]
 [239.91931]
 [241.41238]
 [240.5949 ]
 [240.63133]
 [240.17752]
 [239.48547]
 [239.8201 ]
 [240.56232]
 [239.8804 ]
 [239.43071]
 [239.5325 ]
 [237.14885]
 [242.02179]
 [240.05513]
 [240.09747]
 [242.61647]
 [239.48349]
 [238.00018]
 [241.3304 ]
 [238.57652]
 [240.39828]
 [240.05002]
 [240.23695]
 [242.42032]
 [238.7561 ]
 [241.33542]
 [239.54749]
 [240.51575]
 [240.69147]
 [239.73   ]
 [239.01044]
 [239.38342]
 [240.1422 ]
 [239.95444]
 [238.65465]
 [242.1621 ]
 [238.53963]
 [242.63782]
 [240.91467]
 [242.6879 ]
 [239.44827]
 [240.34888]
 [239.92776]
 [238.52887]
 [241.59381]
 [240.29465]
 [240.21535]
 [239.3244 ]
 [241.55557]
 [239.84467]
 [241.85043]
 [240.26277]
 [240.52608]
 [237.63354]
 [239.6266 ]
 [242.49947]
 [239.36186]
 [239.66986]
 [238.32126]
 [242.05568]
 [238.70213]
 [239.09984]
 [241.84937]
 [240.88976]
 [241.00789]
 [240.52898]
 [240.88861]
 [242.32042]
 [240.81001]
 [239.56715]
 [238.29324]
 [241.9935 ]
 [238.26228]
 [241.90063]
 [239.78397]
 [238.29419]
 [242.43782]
 [241.04893]
 [239.61726]
 [240.46172]
 [239.82378]
 [240.4032 ]
 [241.04828]
 [241.51292]
 [241.8598 ]
 [239.71767]
 [240.92216]
 [237.86807]
 [241.95995]
 [238.2572 ]
 [241.14372]
 [241.92805]
 [241.4039 ]
 [240.45529]
 [241.70468]
 [241.97133]
 [238.57672]
 [239.29973]
 [240.57037]
 [239.79863]
 [239.6932 ]
 [238.05994]
 [242.5538 ]
 [243.31842]
 [238.57735]
 [239.36751]
 [240.9712 ]
 [237.56993]
 [239.04338]
 [241.6832 ]
 [239.04506]
 [238.92732]
 [242.53647]
 [239.77605]
 [240.4215 ]
 [239.36087]
 [240.43205]
 [241.50223]
 [236.72217]
 [240.5569 ]
 [240.62311]
 [241.99855]
 [239.84839]
 [238.43608]
 [238.82683]
 [241.13512]
 [241.19713]
 [240.66077]
 [238.66989]
 [237.70866]
 [242.02312]
 [240.0887 ]
 [240.2601 ]
 [241.81317]
 [238.55876]
 [237.28969]
 [238.49377]
 [238.8463 ]
 [238.06459]
 [240.90962]
 [240.7098 ]
 [240.48155]
 [238.92471]
 [237.44548]
 [240.75986]
 [240.34258]
 [239.43924]
 [241.78275]
 [240.28801]
 [243.02963]
 [240.35019]
 [238.85245]
 [239.32286]
 [240.20818]
 [242.37108]
 [238.31825]
 [241.05171]
 [240.86955]
 [237.96861]
 [238.80934]
 [239.7673 ]
 [240.89035]
 [239.69182]
 [241.60428]
 [239.8303 ]
 [242.81638]
 [238.48378]
 [237.52937]
 [241.46071]
 [241.19954]
 [239.22542]
 [243.09102]
 [236.582  ]
 [238.768  ]
 [241.68643]
 [238.43997]
 [237.90912]
 [239.63383]
 [239.76974]
 [237.4877 ]
 [239.13734]
 [238.40137]
 [240.25343]
 [239.16542]
 [240.07771]
 [242.68001]
 [240.56496]
 [238.68907]
 [240.78679]
 [242.07994]
 [238.39575]
 [242.35059]
 [240.6014 ]
 [239.80896]
 [239.4517 ]
 [239.19212]
 [240.87256]
 [238.23022]
 [240.7845 ]
 [239.68674]
 [237.50812]
 [240.36943]
 [239.68317]
 [240.37854]
 [239.81186]
 [237.80272]
 [240.44778]
 [240.94168]
 [238.03253]
 [238.9612 ]
 [239.39299]
 [239.69238]
 [240.29393]
 [239.20331]
 [239.43102]
 [241.25232]
 [240.6542 ]
 [237.4755 ]
 [238.8073 ]
 [241.44421]
 [238.90984]
 [239.97655]
 [241.61699]
 [240.9133 ]
 [237.88489]
 [238.7468 ]
 [239.32625]
 [241.64708]
 [237.9833 ]
 [240.06932]
 [238.83665]
 [238.45963]
 [240.56013]
 [238.24812]
 [239.05215]
 [239.84084]
 [239.37708]
 [238.60738]
 [240.06418]
 [241.69847]
 [239.54007]
 [240.9536 ]
 [241.81735]
 [240.23357]
 [239.08232]
 [241.72636]
 [236.9735 ]
 [240.80132]
 [241.00528]
 [239.6473 ]
 [241.86029]
 [242.09077]
 [242.75107]
 [241.57422]
 [237.97469]
 [240.00807]
 [238.78389]
 [242.61165]
 [242.1903 ]
 [239.8875 ]
 [240.9395 ]
 [239.05559]
 [238.76031]
 [242.40533]
 [239.59149]
 [240.26802]
 [241.451  ]
 [241.24869]
 [239.2818 ]
 [240.03596]
 [241.53471]
 [242.1644 ]
 [241.54692]
 [240.33528]
 [239.87918]
 [238.75468]
 [240.19211]
 [242.18568]
 [238.37225]
 [237.75046]
 [242.47304]
 [239.65244]
 [240.23428]
 [241.87683]
 [240.77272]
 [239.66182]
 [240.84381]
 [242.81984]
 [240.98723]
 [240.11171]
 [238.65718]
 [241.87737]
 [241.28159]
 [240.68652]
 [239.8653 ]
 [241.469  ]
 [240.54756]
 [237.96756]
 [240.20406]
 [241.27292]
 [239.62927]
 [240.87831]
 [238.76646]
 [241.62102]
 [240.0914 ]
 [241.9476 ]
 [239.69574]
 [238.97153]
 [239.92844]
 [239.18869]
 [238.51332]
 [237.77432]
 [237.41618]
 [241.42969]
 [237.34433]
 [238.09758]
 [240.47769]
 [239.9959 ]
 [240.68018]
 [240.46146]
 [239.30618]
 [239.4498 ]
 [241.60782]
 [239.1716 ]
 [237.33879]
 [239.02942]
 [239.58159]
 [239.40623]
 [240.65163]
 [239.27502]
 [241.41646]
 [239.79617]
 [240.71667]
 [239.89827]
 [237.71616]
 [239.88324]
 [240.29128]
 [238.48344]
 [241.70576]
 [241.4278 ]
 [240.10652]
 [237.98242]
 [238.40457]
 [243.32182]
 [243.3605 ]
 [237.05482]
 [241.278  ]
 [240.66345]
 [239.48485]
 [239.18674]
 [241.66287]
 [238.42543]
 [240.55281]
 [239.55719]
 [241.88202]
 [241.9002 ]
 [239.63548]
 [240.26192]
 [239.50676]
 [240.09077]
 [237.09032]
 [239.32863]
 [237.7806 ]
 [239.64789]
 [241.14462]
 [238.07864]
 [237.54906]
 [240.65332]
 [241.19536]
 [240.67093]
 [239.35612]
 [240.2825 ]
 [242.58885]
 [238.68527]
 [238.35678]
 [242.11064]
 [240.44238]
 [239.63594]
 [242.18167]
 [242.58324]
 [241.0693 ]
 [243.07332]
 [241.13927]
 [240.50433]
 [237.88286]
 [240.83691]
 [239.32355]
 [241.19353]
 [238.8096 ]
 [242.31042]
 [239.65555]
 [242.45224]
 [239.68387]
 [240.35135]
 [237.69102]
 [241.80244]
 [241.7777 ]
 [240.62477]
 [241.46332]
 [240.31718]
 [242.21774]
 [241.49226]
 [238.60243]
 [242.18298]
 [242.98557]
 [241.84073]
 [240.09698]
 [241.98436]
 [240.62288]
 [241.16562]
 [238.0493 ]
 [239.63033]
 [239.83139]
 [239.01561]
 [239.48445]
 [242.44072]
 [239.6133 ]
 [239.25217]
 [240.64374]
 [240.58447]
 [239.51062]
 [240.19933]
 [239.78569]
 [239.63194]
 [240.42274]
 [238.61543]
 [241.03627]
 [240.72021]
 [241.66197]
 [239.21928]
 [240.47607]
 [237.20424]
 [240.1231 ]
 [237.19554]
 [241.86449]
 [239.7897 ]
 [240.4075 ]
 [239.60208]
 [240.79906]
 [240.49599]
 [240.8929 ]
 [237.27519]
 [239.80855]
 [237.72949]
 [239.13757]
 [242.79057]
 [241.90407]
 [243.04485]
 [239.50716]
 [242.64618]
 [239.76901]
 [238.26093]
 [240.59685]
 [236.6914 ]
 [241.71785]
 [238.8095 ]
 [241.61716]
 [240.25237]
 [237.92387]
 [240.58522]
 [242.7438 ]
 [241.63629]
 [238.64319]
 [239.41666]
 [239.84274]
 [241.623  ]
 [241.49005]
 [242.06618]
 [239.98228]
 [241.3046 ]
 [241.00963]
 [240.19084]
 [240.74382]
 [238.78204]
 [240.62173]
 [239.34108]
 [238.33295]
 [240.29967]
 [237.58182]
 [239.66318]
 [239.19998]
 [241.2582 ]
 [238.02853]
 [240.16626]
 [239.64803]
 [241.99292]
 [240.06166]
 [241.65771]
 [239.96002]
 [241.15987]
 [239.72426]
 [239.9587 ]
 [240.03816]
 [240.00177]
 [241.07993]
 [238.41414]
 [239.7907 ]
 [240.239  ]
 [240.08475]
 [241.39331]
 [241.34552]
 [238.28522]
 [239.497  ]
 [240.21884]
 [241.31274]
 [240.57562]
 [238.03311]
 [238.75092]
 [240.81784]
 [241.92813]
 [237.29222]
 [238.03   ]
 [243.0442 ]
 [239.76521]
 [241.42638]
 [238.50804]
 [239.3811 ]
 [241.8099 ]
 [242.77568]
 [237.32445]
 [240.84239]
 [239.49191]
 [240.28244]
 [241.94148]
 [238.64024]
 [241.73065]
 [242.40482]
 [242.41382]
 [242.24121]
 [240.84077]
 [239.86383]
 [240.77966]
 [239.63887]
 [241.7363 ]
 [239.1438 ]
 [239.32867]
 [240.61522]
 [242.47191]
 [240.33727]
 [238.93947]
 [242.21051]
 [242.16553]
 [240.5177 ]
 [238.83617]
 [240.83719]
 [240.06104]
 [240.1964 ]
 [240.0481 ]
 [237.72366]
 [240.11127]
 [238.8766 ]
 [238.10135]
 [239.71742]
 [238.11139]
 [242.13354]
 [242.51416]
 [240.26059]
 [240.47868]
 [240.22078]
 [241.42374]
 [240.19022]
 [240.35313]
 [243.18138]
 [240.62067]
 [240.57301]
 [238.9564 ]
 [240.36795]
 [240.2156 ]
 [240.18488]
 [237.94504]
 [240.65126]
 [239.83638]
 [240.43568]
 [240.6386 ]
 [242.18245]
 [239.39854]
 [238.62038]
 [238.70534]
 [241.06924]
 [238.75412]
 [241.45529]
 [240.87114]
 [240.01608]
 [237.17065]
 [241.08356]
 [239.47484]
 [240.13141]
 [238.2204 ]
 [239.15114]
 [239.573  ]
 [236.01627]
 [241.07277]
 [241.5825 ]
 [239.13278]
 [243.26369]
 [240.3767 ]
 [237.43753]
 [237.88583]
 [240.70876]
 [239.20157]
 [239.61928]
 [239.79228]
 [242.03522]
 [241.05937]
 [240.18494]
 [239.93652]
 [239.11913]
 [241.68521]
 [239.16136]
 [240.05011]
 [238.51146]
 [237.16516]
 [239.11482]
 [241.99185]
 [238.80852]
 [238.7958 ]
 [241.13753]
 [239.6482 ]
 [239.55832]
 [239.23395]
 [240.30563]
 [238.92003]
 [241.28403]
 [241.73755]
 [238.17686]
 [239.97328]
 [242.26729]
 [237.1443 ]
 [238.585  ]
 [239.66739]
 [241.88637]
 [242.04256]
 [242.16724]
 [238.49408]
 [242.60172]
 [236.15596]
 [241.57124]
 [237.99713]
 [238.2664 ]
 [240.26488]
 [241.64207]
 [238.27486]]


predict_X = model.predict(np.expand_dims(X_train[-1], axis=0))

1/1 [==============================] - 0s 28ms/step


# Reshape predict_X to have a shape of (1, 1)
predict_X = predict_X.reshape(1, 1)

# Inverse transform predictions to original scale
predict_X = scaler.inverse_transform(predict_X)


print(predict_X)

[[238.27486]]


# Create a copy of X_test to store the updated values
X_test_updated = X_train[-1].reshape(-1, 1)
num_days_to_predict = len(test)
print(X_test_updated.shape)

(10, 1)


# Make predictions for the next 90 days and update X_test
predictions_list = []
for i in range(num_days_to_predict):
    # Make predictions for the current day
    predictions_new = model.predict(X_test_updated)
    #predictions = scaler.inverse_transform(predictions)

    # Append the predictions to X_test_updated for the next day
    X_test_updated = np.append(X_test_updated, predictions_new, axis=1)

    # Delete the first element from the array
    X_test_updated = X_test_updated[:, 1:]

    # Append the predictions to the list for plotting later
    predictions_list.append(predictions_new[0, 0])  # Assuming the prediction is a single value (e.g., the first column)

# reshape the predictions_list
predictions_list = np.array(predictions_list).reshape(-1, 1)
# apply the scaler.inverse_transform functio nto all the values in the predictions_list
predictions_list = scaler.inverse_transform(predictions_list)
# make a data frame from the predictions_list and add a column based on df['Date'][252:]
df_predictions = pd.DataFrame(predictions_list, columns=['Price'])

1/1 [==============================] - 0s 479ms/step
1/1 [==============================] - 0s 28ms/step
1/1 [==============================] - 0s 27ms/step
1/1 [==============================] - 0s 29ms/step
1/1 [==============================] - 0s 30ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 29ms/step
1/1 [==============================] - 0s 36ms/step
1/1 [==============================] - 0s 29ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 27ms/step
1/1 [==============================] - 0s 24ms/step
1/1 [==============================] - 0s 30ms/step
1/1 [==============================] - 0s 31ms/step
1/1 [==============================] - 0s 28ms/step
1/1 [==============================] - 0s 24ms/step
1/1 [==============================] - 0s 29ms/step
1/1 [==============================] - 0s 28ms/step
1/1 [==============================] - 0s 27ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 29ms/step
1/1 [==============================] - 0s 27ms/step
1/1 [==============================] - 0s 24ms/step
1/1 [==============================] - 0s 24ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 24ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 30ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 28ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 28ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 27ms/step
1/1 [==============================] - 0s 31ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 33ms/step
1/1 [==============================] - 0s 32ms/step
1/1 [==============================] - 0s 23ms/step
1/1 [==============================] - 0s 25ms/step
1/1 [==============================] - 0s 29ms/step
1/1 [==============================] - 0s 23ms/step
1/1 [==============================] - 0s 23ms/step
1/1 [==============================] - 0s 32ms/step
1/1 [==============================] - 0s 29ms/step
1/1 [==============================] - 0s 28ms/step
1/1 [==============================] - 0s 26ms/step
1/1 [==============================] - 0s 36ms/step
1/1 [==============================] - 0s 30ms/step
1/1 [==============================] - 0s 30ms/step
1/1 [==============================] - 0s 33ms/step
1/1 [==============================] - 0s 31ms/step
1/1 [==============================] - 0s 30ms/step


# Filter the dates to keep only those from the year 2021
dates_2021 = test['Date']

# Reset the index to start from 0
dates_2021.reset_index(drop=True, inplace=True)

# Assign the filtered dates to the 'Date' column in 'df_predictions'
df_predictions['Date'] = dates_2021

df_predictions = df_predictions[['Date','Price']]


# plot the predictions_list
plt.plot(df_pd['Date'], df_pd['Price'], label ='Actual Price')
plt.tick_params(axis='x', rotation=45)
plt.plot(df_predictions['Date'],df_predictions['Price'], label = 'Predicted Price')
plt.legend()
plt.show()


# Define the window size
window = 10

# Concatenate the last window of the original data with the predictions
df_predictions_Bolinger = pd.concat([train[['Date','Price']].tail(window),df_predictions]).reset_index(drop=True)


df_predictions_Bolinger


# Calculate Bollinger Bands
def calculate_bollinger_bands(db, window):
    db['SMA'] = db['Price'].rolling(window=window).mean()
    db['STD'] = db['Price'].rolling(window=window).std()
    db['High'] = db['SMA'] + (2 * db['STD'])
    db['Low'] = db['SMA'] - (2 * db['STD'])
    return db

window_size = 20

# Visualize the stock prices and Bollinger Bands
df_b = calculate_bollinger_bands(df_predictions_Bolinger, window_size)

plt.figure(figsize=(16, 8))
plt.plot(df_pd['Date'], df_pd['Price'], label='Stock Price')
plt.plot(df_b['Date'], df_b['High'], label='Upper Bollinger Band')
plt.plot(df_b['Date'], df_b['Low'], label='Lower Bollinger Band')
plt.plot(df_predictions['Date'], df_predictions['Price'],label='Predcited')
plt.title('Stock Price and Bollinger Bands')
plt.xlabel('Date')
plt.ylabel('Price')
plt.legend()
plt.show()


test = test.reset_index(drop=True)


df_b = df_b.tail(len(test)).reset_index(drop=True)


df_b['Signal'] = ''


for i in range(len(test)):
  if test['Price'][i] < df_b['Low'][i]:
    df_b['Signal'][i] = 'BUY'
  elif test['Price'][i] > df_b['Low'][i] and test['Price'][i] < df_b['High'][i]:
    df_b['Signal'][i] = 'HOLD'
  else: df_b['Signal'][i] = 'Sell'


# Visualize the results
plt.figure(figsize=(16, 8))
plt.plot(test['Date'], test['Price'], label='Stock Price')
plt.plot(df_b['Date'].tail(len(test)), df_b['High'].tail(len(test)), label='Upper Bollinger Band')
plt.plot(df_b['Date'].tail(len(test)), df_b['Low'].tail(len(test)), label='Lower Bollinger Band')

# plt.plot(df_predictions['Date'], df_predictions['Price'],label='Predcited')
plt.title('Stock Price and Bollinger Bands for the Q1 2021')
plt.xlabel('Date')
plt.ylabel('Price')
plt.legend()
plt.show()

# Display buy/sell signals
print(df_b['Signal'].value_counts())

Signal
Sell    54
HOLD     7
Name: count, dtype: int64

	Price	Time_Stamp
count	311.000000	311
mean	234.524019	2020-08-15 01:23:20.643086848
min	174.270000	2020-01-03 00:00:00
25%	207.700000	2020-04-23 12:00:00
50%	229.140000	2020-08-13 00:00:00
75%	264.380000	2020-12-05 12:00:00
max	293.300000	2021-03-30 00:00:00
std	32.079847	NaN

	Date	Price	Open	High	Low	Month	Year
Time_Stamp
2021-01-29	Feb 01, 2021	263.8	260	264.06	258.73	1	2021
2021-02-01	Feb 02, 2021	263.15	265	268.39	263.05	2	2021
2021-02-02	Feb 03, 2021	263.53	264.32	268.14	261.05	2	2021
2021-02-03	Feb 04, 2021	270.11	263.21	270.6	262.34	2	2021
2021-02-04	Feb 05, 2021	271.7	271.99	272.57	268.87	2	2021
2021-02-05	Feb 08, 2021	275.2	273.01	275.58	272.32	2	2021
2021-02-08	Feb 09, 2021	268.98	275.99	276.29	267.19	2	2021
2021-02-09	Feb 10, 2021	265.9	269.28	270.99	265.2	2	2021
2021-02-10	Feb 11, 2021	264.7	265.5	266.9	263.75	2	2021
2021-02-11	Feb 12, 2021	266	264	266.44	258.55	2	2021
2021-02-12	Feb 15, 2021	271.81	268.49	271.87	267.07	2	2021
2021-02-16	Feb 16, 2021	270.9	272.3	272.8	268.27	2	2021
2021-02-17	Feb 17, 2021	270.9	270.99	275.07	267.56	2	2021
2021-02-18	Feb 18, 2021	266.99	270.35	272.46	266.21	2	2021
2021-02-19	Feb 19, 2021	271.34	266.12	272.81	264.19	2	2021
2021-02-22	Feb 20, 2021	270.51	271	271	269.02	2	2021
2021-02-23	Feb 22, 2021	273.71	270.1	274.46	268.12	2	2021
2021-02-24	Feb 24, 2021	271.9	272.12	272.75	269.42	2	2021
2021-02-25	Feb 25, 2021	270.71	273.6	276.48	270.65	2	2021
2021-02-26	Feb 26, 2021	270.17	268	272.49	266.43	2	2021
2021-03-01	Mar 01, 2021	273.58	273	274.73	271.13	3	2021
2021-03-02	Mar 02, 2021	279.1	271.89	280	271.6	3	2021
2021-03-03	Mar 03, 2021	278.01	281.7	283.83	277.63	3	2021
2021-03-04	Mar 04, 2021	274.75	278.5	281.96	273.63	3	2021
2021-03-05	Mar 05, 2021	277.5	276	278.98	272.76	3	2021
2021-03-08	Mar 09, 2021	285.79	278.8	286.38	277.56	3	2021
2021-03-09	Mar 10, 2021	282.24	285.24	288	281	3	2021
2021-03-10	Mar 11, 2021	286	283.84	286.38	282.3	3	2021
2021-03-11	Mar 12, 2021	284.93	286.3	287.49	282.82	3	2021
2021-03-12	Mar 15, 2021	292.74	286.05	292.92	285.28	3	2021
2021-03-15	Mar 16, 2021	292.33	293.06	293.88	288.61	3	2021
2021-03-16	Mar 17, 2021	283.35	292.35	293.25	279.6	3	2021
2021-03-17	Mar 18, 2021	279.25	284.04	285.79	278.24	3	2021
2021-03-18	Mar 19, 2021	285.47	279	286.31	278.03	3	2021
2021-03-19	Mar 22, 2021	288.65	285	291.2	283.08	3	2021
2021-03-22	Mar 23, 2021	282.6	287.56	288.3	282.11	3	2021
2021-03-23	Mar 24, 2021	285.79	283.4	288	282.25	3	2021
2021-03-24	Mar 25, 2021	286.66	286.5	287.03	283.85	3	2021
2021-03-25	Mar 26, 2021	291.22	288.63	292.75	288.32	3	2021
2021-03-26	Mar 29, 2021	293.3	289.72	294.09	289.26	3	2021
2021-03-29	Mar 30, 2021	293.25	294.86	295.72	291.5	3	2021
2021-03-30	Mar 31, 2021	291.02	294	294.42	290.26	3	2021

	Price
Time_Stamp
2021-01-22	265.3
2021-01-25	269.34
2021-01-26	265.5
2021-01-27	265.96
2021-01-28	258.11

	Date	Price	Open	High	Low	Month	Year
Time_Stamp
2021-01-22	Jan 25, 2021	265.3	271	273.63	261.65	1	2021
2021-01-25	Jan 26, 2021	269.34	262.7	269.57	262.61	1	2021
2021-01-26	Jan 27, 2021	265.5	270.07	273.8	265.02	1	2021
2021-01-27	Jan 28, 2021	265.96	263.5	267.17	261.17	1	2021
2021-01-28	Jan 29, 2021	258.11	263.77	264.75	257.36	1	2021

	Date	Price	Open	High	Low	Month	Year
Time_Stamp
2021-01-29	Feb 01, 2021	263.8	260	264.06	258.73	1	2021
2021-02-01	Feb 02, 2021	263.15	265	268.39	263.05	2	2021
2021-02-02	Feb 03, 2021	263.53	264.32	268.14	261.05	2	2021
2021-02-03	Feb 04, 2021	270.11	263.21	270.6	262.34	2	2021
2021-02-04	Feb 05, 2021	271.7	271.99	272.57	268.87	2	2021

ValueInvestor¶

Plot the Time Series to understand the behaviour of the data.¶

Check the basic measures of descriptive statistics¶

Plot the boxplot of the 'Open' variable with respect to the different years.¶

Plot the boxplot of the 'Open' variable with respect to months of different.¶

Decompose the 'Price' variable Time Series.¶

Additive Decomposition¶

Multiplicative Decomposition¶

Split the data into training and test set before building the models.¶

Check for stationarity of the series at $\alpha$ = 0.05 using appropriate statistical tests.¶

Series is stationary at 5% Alpha as per Dicky Fuller¶

Model Selection¶

Build an Automated version of an ARIMA model for which the best parameters are selected in accordance with the lowest Akaike Information Criteria (AIC).¶

Note: The data has some seasonality so ideally we should build a SARIMA model. But for demonstration purposes we are building an ARIMA model both by looking at the minimum AIC criterion and by looking at the ACF and the PACF plots.¶

Predict on the Test Set using this model and evaluate the model.¶

Build a version of the ARIMA model for which the best parameters are selected by looking at the ACF and the PACF plots.¶

Predict on the Test Set using this model and evaluate the model.¶

Build an Automated version of a SARIMA model for which the best parameters are selected in accordance with the lowest Akaike Information Criteria (AIC).¶

Predict on the Test Set using this model and evaluate the model.¶

Build a version of the SARIMA model for which the best parameters are selected by looking at the ACF and the PACF plots. - Seasonality at 6.¶

Predict on the Test Set using this model and evaluate the model.¶

Building the most optimum model on the Full Data.¶

Note: If we have a scenario where our training data was stationary but our full data was not stationary. We can still use the same parameters as our training data but with adding an appropriate level of differencing or transformation which is needed for the data to be stationary.¶

Evaluate the model on the whole data and predict 20 months into the future (till the end of next year).¶

Build Time Series models using the Prophet Library.¶

Adding weekends¶

Applying LSTM Model¶

Plotting Current Prices of year 2020¶

LSTM Data preparation¶

Training LSTM Model¶

Plotting Actual vs Forecasted for the 1st Quarter 2021¶

Calculating Bolinger for LSTM¶

Buy/Hold/Sell Strategy¶

Conclusion¶

	Date	Price	Open	High	Low
Time_Stamp
2020-01-03	Jan 03, 2020	255	255.99	258.19	253.7
2020-01-06	Jan 06, 2020	253.9	254.75	254.84	251.4
2020-01-07	Jan 08, 2020	259.15	253.57	259.15	253.03
2020-01-08	Jan 09, 2020	257.99	259.4	261.76	257.01
2020-01-09	Jan 10, 2020	258.19	257.86	259.25	256.88
...	...	...	...	...	...
2021-03-24	Mar 25, 2021	286.66	286.5	287.03	283.85
2021-03-25	Mar 26, 2021	291.22	288.63	292.75	288.32
2021-03-26	Mar 29, 2021	293.3	289.72	294.09	289.26
2021-03-29	Mar 30, 2021	293.25	294.86	295.72	291.5
2021-03-30	Mar 31, 2021	291.02	294	294.42	290.26

	param	AIC
17	(2, 1, 2)	1847.229343
3	(0, 1, 0)	1849.818816
4	(0, 1, 1)	1851.625113
9	(1, 1, 0)	1851.644078
15	(2, 1, 0)	1852.857987

Price	mean	mean_se	mean_ci_lower	mean_ci_upper
311	255.237032	4.990792	245.455260	265.018805
312	253.711486	7.186373	239.626453	267.796519
313	255.364587	8.853243	238.012549	272.716625
314	255.224724	10.252391	235.130408	275.319041
315	256.306100	11.482296	233.801214	278.810986

	RMSE	MAPE
ARIMA(2,1,2)	24.676528	8.141942
ARIMA(3,1,3)	23.750883	7.844811
SARIMA(1,1,3)(3,0,3,6)	19.783019	6.583532

Price	mean	mean_se	mean_ci_lower	mean_ci_upper
311	290.728823	4.544049	281.822650	299.634995
312	293.274472	6.691808	280.158770	306.390174
313	291.228576	8.238231	275.081941	307.375211
314	291.340441	9.253123	273.204654	309.476229
315	291.861067	10.288893	271.695208	312.026927

	ds	y	High	Low
0	2020-01-03	255.00	258.19	253.7
1	2020-01-06	253.90	254.84	251.4
2	2020-01-08	259.15	259.15	253.03
3	2020-01-09	257.99	261.76	257.01
4	2020-01-10	258.19	259.25	256.88

	ds	trend	yhat_lower	yhat_upper	trend_lower	trend_upper	Christmas Day	Christmas Day_lower	Christmas Day_upper	Christmas Day (observed)	...	holidays	holidays_lower	holidays_upper	weekly	weekly_lower	weekly_upper	multiplicative_terms	multiplicative_terms_lower	multiplicative_terms_upper	yhat
0	2020-01-03	274.464274	263.277074	288.906750	274.464274	274.464274	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	2.098955	2.098955	2.098955	0.0	0.0	0.0	276.563229
1	2020-01-06	271.597898	260.548601	287.485340	271.597898	271.597898	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	2.622925	2.622925	2.622925	0.0	0.0	0.0	274.220823
2	2020-01-08	269.686980	260.015063	285.818562	269.686980	269.686980	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	4.020814	4.020814	4.020814	0.0	0.0	0.0	273.707794
3	2020-01-09	268.731522	258.792743	284.989911	268.731522	268.731522	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	2.645015	2.645015	2.645015	0.0	0.0	0.0	271.376536
4	2020-01-10	267.776063	257.006874	283.698388	267.776063	267.776063	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	2.098955	2.098955	2.098955	0.0	0.0	0.0	269.875018
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
349	2021-05-25	292.672295	279.355837	313.488693	281.306348	304.033579	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	3.998956	3.998956	3.998956	0.0	0.0	0.0	296.671251
350	2021-05-26	292.856431	279.057604	315.759033	281.089633	304.720885	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	4.020814	4.020814	4.020814	0.0	0.0	0.0	296.877245
351	2021-05-27	293.040567	277.478037	313.936166	280.819845	305.374246	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	2.645015	2.645015	2.645015	0.0	0.0	0.0	295.685581
352	2021-05-28	293.224702	277.586141	313.999221	280.542868	306.204868	0.0	0.0	0.0	0.0	...	0.000000	0.000000	0.000000	2.098955	2.098955	2.098955	0.0	0.0	0.0	295.323657
353	2021-05-31	293.777109	268.643621	305.391002	280.642544	307.089386	0.0	0.0	0.0	0.0	...	-8.966533	-8.966533	-8.966533	2.622925	2.622925	2.622925	0.0	0.0	0.0	287.433502

	ds	yhat	yhat_lower	yhat_upper	y	cutoff
0	2020-07-31	220.807567	218.027898	223.860076	221.57	2020-07-30
1	2020-08-03	225.917434	222.782095	229.185630	226.72	2020-07-30
2	2020-08-04	227.245497	224.334669	230.234302	227.16	2020-07-30
3	2020-08-05	227.774604	224.818455	230.863083	226.40	2020-07-30
4	2020-08-06	226.285449	223.317876	229.482484	226.00	2020-07-30
...	...	...	...	...	...	...
264	2021-03-24	284.238138	281.650176	287.146387	285.79	2021-02-15
265	2021-03-25	283.995201	281.230699	286.996797	286.66	2021-02-15
266	2021-03-26	289.615340	286.839556	292.696044	291.22	2021-02-15
267	2021-03-29	290.679844	287.898022	293.406249	293.30	2021-02-15
268	2021-03-30	292.134173	289.382241	295.071645	293.25	2021-02-15

	Date	Price	Open	High	Low	Time_Stamp
250	2021-01-04	272.00	274.67	276.6	270.61	2020-12-31
251	2021-01-05	275.20	272.69	275.7	270.28	2021-01-04
252	2021-01-06	272.93	275.94	275.94	271.62	2021-01-05
253	2021-01-08	283.64	275.11	283.99	275.11	2021-01-06
254	2021-01-11	289.49	282.86	292.95	279.48	2021-01-07
...	...	...	...	...	...	...
306	2021-03-25	286.66	286.5	287.03	283.85	2021-03-24
307	2021-03-26	291.22	288.63	292.75	288.32	2021-03-25
308	2021-03-29	293.30	289.72	294.09	289.26	2021-03-26
309	2021-03-30	293.25	294.86	295.72	291.5	2021-03-29
310	2021-03-31	291.02	294	294.42	290.26	2021-03-30

	Date	Price
0	2020-12-17	278.500000
1	2020-12-18	270.160000
2	2020-12-21	261.790000
3	2020-12-22	262.040000
4	2020-12-23	270.100000
...	...	...
66	2021-03-25	209.040359
67	2021-03-26	209.040359
68	2021-03-29	209.040359
69	2021-03-30	209.040359
70	2021-03-31	209.040359